List of large language models

This is a dynamic list and may never be able to satisfy particular standards for completeness. You can help by editing the page to add missing items, with references to reliable sources.

A large language model (LLM) is a type of machine learning model designed for natural language processing tasks such as language generation. LLMs are language models with many parameters, and are trained with self-supervised learning on a vast amount of text.

List

For the training cost column, 1 petaFLOP-day equals 1 petaFLOP/sec × 1 day, or 8.64×1019 FLOP (floating point operations). Only the cost of the largest model is shown. The number of parameters is measured in billions, and the training cost is measured in petaFLOP-days.

2018

Name Release date Developer # of params Corpus size Training cost License Notes
GPT-1 Jun 11, 2018 OpenAI 0.117B Unknown 1 MIT
First GPT model, decoder-only transformer. Trained for 30 days on 8 P600 GPUs.
BERT Oct 2018 Google 0.340B 3.3B words 9 Apache 2.0
An early and influential language model. Encoder-only and thus not built to be prompted or generative. Training took 4 days on 64 TPUv2 chips.

2019

Name Release date Developer # of params Corpus size Training cost License Notes
T5 Oct 2019 Google 11B 34B tokens Unknown Apache 2.0
Base model for Google projects like Imagen.
XLNet Jun 2019 Google 0.340B 33B words 330 Apache 2.0
An alternative to BERT; designed as encoder-only. Trained on 512 TPU v3 chips for 5.5 days.
GPT-2 Feb 2019 OpenAI 1.5B 40GB (~10B tokens) 28 MIT
Trained on 32 TPUv3 chips for 1 week.

2020

Name Release date Developer # of params Corpus size Training cost License Notes
GPT-3 May 2020 OpenAI 175B 300B tokens 3640 Proprietary
A fine-tuned variant of GPT-3, termed GPT-3.5, was made available to the public through ChatGPT in 2022.

2021

Name Release date Developer # of params Corpus size Training cost License Notes
GPT-Neo Mar 2021 EleutherAI 2.7B 825 GiB Unknown MIT
The first of a series of free GPT-3 alternatives released by EleutherAI. GPT-Neo outperformed an equivalent-size GPT-3 model on some benchmarks, but was significantly worse than the largest GPT-3.
HyperCLOVA May 2021 Naver 204B
82B		 560B Tokens Unknown Unreleased
Its technical report says the model is the "first discovery on near 100B-scale non-English LM" by "constructing a large Korean-centric corpus of 560B tokens."
GPT-J Jun 2021 EleutherAI 6B 825 GiB 200 Apache 2.0
Megatron-Turing NLG Oct 2021 Microsoft and Nvidia 530B 338.6B tokens 38000 Unreleased
Trained for 3 months on over 2000 A100 GPUs on the NVIDIA Selene Supercomputer, for over 3 million GPU-hours.
Ernie 3.0 Titan Dec 2021 Baidu 260B 4TB Unknown Proprietary
Claude Dec 2021 Anthropic 52B 400B tokens Unknown
Fine-tuned for desirable behavior in conversations.
GLaM (Generalist Language Model) Dec 2021 Google 1162B 1.6T tokens 5600
Gopher Dec 2021 Google DeepMind 280B 300B tokens 5833

2022

Name Release date Developer # of params Corpus size Training cost License Notes
LaMDA (Language Models for Dialog Applications) Jan 2022 Google 137B 1.56T words, 168B tokens 4110 Proprietary
GPT-NeoX Feb 2022 EleutherAI 20B 825 GiB 740 Apache 2.0
Chinchilla scaling Mar 2022 Google DeepMind 70B 1.4T tokens 6805 Proprietary
PaLM (Pathways Language Model) Apr 2022 Google 540B 768B tokens 29,250
Trained for ~60 days on ~6000 TPU v4 chips.
OPT (Open Pretrained Transformer) May 2022 Meta 175B 180B tokens 310 Non-commercial research
GPT-3 architecture with some adaptations from Megatron. The training logbook written by the team was published. The smaller models including 66B are publicly available, while the 175B model is available on request.
YaLM 100B Jun 2022 Yandex 100B 1.7TB Unknown Apache 2.0
Minerva Jun 2022 Google 540B 38.5B tokens from webpages filtered for math content and from arXiv Proprietary
For solving "mathematical and scientific questions using step-by-step reasoning".
BLOOM Jul 2022 Large collaboration led by Hugging Face 175B 350B tokens (1.6TB) Responsible AI
Galactica Nov 2022 Meta 120B 106B tokens CC-BY-NC-4.0
AlexaTM (Teacher Models) Nov 2022 Amazon 20B 1.3T Proprietary

2023

Name Release date Developer # of params Corpus size Training cost License Notes
Llama Feb 2023 Meta AI 65B 1.4T 6300 Non-commercial research
Facebook's license and distribution scheme restricted access to approved researchers, but the model weights were leaked and became widely available.
Stanford Alpaca Mar 13, 2023 Stanford University 7B 52K demonstrations <US$100 Non-commercial
Fine-tuned from LLaMA 7B on instruction-following demonstrations generated using OpenAI's text-davinci-003.
GPT-4 Mar 2023 OpenAI Unknown Unknown Unknown Proprietary
Cerebras-GPT Mar 2023 Cerebras 13B 270 Apache 2.0
Falcon Mar 2023 Technology Innovation Institute 40B 1T tokens, from RefinedWeb (filtered web text corpus) plus some "curated corpora". 2800 Apache 2.0
BloombergGPT Mar 2023 Bloomberg L.P. 50B 363B tokens from Bloomberg's proprietary data sources, plus 345B tokens from general purpose datasets Unknown Unreleased
Designed for financial tasks.
PanGu-Σ Mar 2023 Huawei 1085B 329B tokens Proprietary
OpenAssistant Mar 2023 LAION 17B 1.5T tokens Apache 2.0
Vicuna Mar 2023 LMSYS 13B 70K conversations US$300 Non-commercial
Fine-tuned from LLaMA on user-shared conversations collected from ShareGPT.
Jurassic-2 Mar 2023 AI21 Labs Unknown Unknown Unknown Proprietary
PaLM 2 (Pathways Language Model 2) May 2023 Google 340B 3.6T tokens 85,000
Used in the Bard chatbot.
YandexGPT May 17, 2023 Yandex Unknown Unknown Unknown
Phi-1 Jun 21, 2023 Microsoft 1.3B 7B tokens Unknown MIT
Trained for 4 days on 8 A100s.
Llama 2 Jul 2023 Meta AI 70B 2T tokens 21,000 Llama 2
Trained over 3.3 million GPU (A100) hours.
Claude 2 Jul 2023 Anthropic Unknown Unknown Unknown Proprietary
Used in the Claude chatbot.
Granite 13b Jul 2023 IBM Unknown
Used in IBM Watsonx.
Mistral 7B Sep 2023 Mistral AI 7.3B Apache 2.0
YandexGPT 2 Sep 7, 2023 Yandex Unknown Proprietary
Claude 2.1 Nov 2023 Anthropic Unknown
Used in the Claude chatbot. Has a context window of 200,000 tokens, or ~500 pages.
Grok-1 Nov 2023 xAI 314B Apache 2.0
Used in the Grok chatbot. Grok 1 has a context length of 8,192 tokens and has access to X (Twitter).
Gemini 1.0 Dec 2023 Google DeepMind Unknown Proprietary
Multimodal model, comes in three sizes. Used in the chatbot of the same name.
Mixtral 8x7B Dec 2023 Mistral AI 46.7B Apache 2.0
Outperforms GPT-3.5 and Llama 2 70B on many benchmarks. Mixture of experts model, with 12.9 billion parameters activated per token.
DeepSeek-LLM Nov 29, 2023 DeepSeek 67B 2T tokens: table 2  12,000 DeepSeek
Trained on English and Chinese text. Used 1024 training FLOPs for 67B model, 10b FLOPs for 7B.: figure 5 
Phi-2 Dec 2023 Microsoft 2.7B 1.4T tokens 419 MIT
Trained on real and synthetic "textbook-quality" data over 14 days on 96 A100 GPUs.

2024

Name Release date Developer # of params Corpus size Training cost License Notes
Gemini 1.5 Feb 2024 Google DeepMind Unknown Unknown Unknown Proprietary
Multimodal model based on a MoE architecture. Context window above 1 million tokens.
Gemini Ultra Feb 2024
Gemma Feb 2024 7B 6T tokens Gemma Terms of Use
OLMo Feb 2024 Allen Institute for AI 7B 2T tokens Apache 2.0
Claude 3 Mar 2024 Anthropic Unknown Unknown Proprietary
Includes three models: Haiku, Sonnet, and Opus.
DBRX Mar 2024 Databricks and Mosaic ML 136B 12T tokens Databricks Open Model
YandexGPT 3 Pro Mar 28, 2024 Yandex Unknown Unknown Proprietary
Fugaku-LLM May 2024 Fujitsu, Tokyo Institute of Technology, Tohoku University, RIKEN, etc. 13B 380B tokens Fugaku-LLM Terms of Use
The largest model ever trained on CPU-only, on the Fugaku supercomputer; the model was trained from scratch on 380 billion tokens using 13,824 Fugaku nodes.
Chameleon May 2024 Meta AI 34B 4.4T Non-commercial research
Mixtral 8x22B Apr 17, 2024 Mistral AI 141B Unknown Apache 2.0
Phi-3 Apr 23, 2024 Microsoft 14B 4.8T tokens MIT
Marketed by Microsoft as a "small language model".
Granite Code Models May 2024 IBM Unknown Unknown Apache 2.0
YandexGPT 3 Lite May 28, 2024 Yandex Unknown Unknown Proprietary
Qwen2 Jun 2024 Alibaba Cloud 72B 3T tokens Various
DeepSeek-V2 Jun 2024 DeepSeek 236B 8.1T tokens 28,000 DeepSeek
1.4M hours on H800.
Nemotron-4 Jun 2024 Nvidia 340B 9T tokens 200,000 NVIDIA Open Model
Trained for 1 epoch. Trained on 6144 H100 GPUs between December 2023 and May 2024.
Claude 3.5 Jun 2024 Anthropic Unknown Unknown Unknown Proprietary
Initially, only one model, Sonnet, was released. In October 2024, Sonnet 3.5 was upgraded, and Haiku 3.5 became available.
Llama 3.1 Jul 2024 Meta AI 405B 15.6T tokens 440,000 Llama 3
405B version took 31 million hours on H100-80GB, at 3.8E25 FLOPs.
Grok-2 Aug 14, 2024 xAI Unknown Unknown Unknown xAI Community License Agreement
Originally closed-source, then re-released as "Grok 2.5" under a source-available license in August 2025.
OpenAI o1 Sep 12, 2024 OpenAI Unknown Unknown Proprietary
First LLM described as a "reasoning model".[better source needed]
Sarvam-1 Oct 24, 2024 Sarvam AI 2B ~2T tokens Sarvam AI Research
Supports 10 Indic languages and English
YandexGPT 4 Lite and Pro Oct 24, 2024 Yandex Unknown Unknown Proprietary
Mistral Large Nov 2024 Mistral AI 123B Unknown Mistral Research
Upgraded over time. The latest version is 24.11.
Pixtral
Multimodal. There is also a 12B version which is under Apache 2 license.
OLMo 2 Nov 2024 Allen Institute for AI 32B 6.6T tokens 15,000 Apache 2.0
Phi-4 Dec 12, 2024 Microsoft 14B 9.8T tokens Unknown MIT
Marketed by Microsoft as a "small language model".
DeepSeek-V3 Dec 2024 DeepSeek 671B 14.8T tokens 56,000
Used 2.788M training hours on H800 GPUs. Originally released under the DeepSeek License, then re-released under the MIT License as "DeepSeek-V3-0324" in March 2025.
Amazon Nova Dec 2024 Amazon Unknown Unknown Unknown Proprietary
Includes three models: Nova Micro, Nova Lite, and Nova Pro.

2025

Name Release date Developer # of params Corpus size License Notes
DeepSeek-R1 Jan 20 DeepSeek 671B Unknown MIT
No pretraining; reinforcement-learned upon V3-Base.
Qwen2.5 Jan 26 Alibaba 72B 18T Various
7 dense models with parameter counts from 0.5B to 72B. Alibaba also released 2 MoE variants.
MiniMax-Text-01 Jan 14 Minimax 456B 4.7T Minimax Model
Gemini 2.0 Feb 5 Google DeepMind Unknown Unknown Proprietary
Three models released: Flash, Flash-Lite and Pro.
Grok 3 Feb 19 xAI
Training cost claimed to be "10x the compute of previous state-of-the-art models".
Claude 3.7 Feb 24 Anthropic
One model, Sonnet 3.7.
YandexGPT 5 Lite Pretrain and Pro Feb 25 Yandex
GPT-4.5 Feb 27 OpenAI
OpenAI's largest non-reasoning model at the time.
Gemini 2.5 Mar 25 Google DeepMind
Three models released: Flash, Flash-Lite and Pro.
YandexGPT 5 Lite Instruct Mar 31 Yandex
Llama 4 Apr 5 Meta AI 400B 40T Llama 4
OpenAI o3 and o4-mini Apr 16 OpenAI Unknown Unknown Proprietary
Reasoning models.
Qwen3 Apr 28 Alibaba Cloud 235B 36T Apache 2.0
Multiple sizes, the smallest being 0.6B.
Claude 4 May 22 Anthropic Unknown Unknown Proprietary
Includes two models, Sonnet and Opus.
Sarvam-M May 23 Sarvam AI 24B Unknown Apache 2.0
Hybrid reasoning model fine-tuned on Mistral Small base; optimized for math, programming, and Indian languages.
Grok 4 Jul 9 xAI Unknown Unknown Proprietary
Param-1 Jul 21 BharatGen 2.9B 5T Apache 2.0
Corpus is "focus[ed] on India’s linguistic landscape".
GLM-4.5 Jul 29 Z.ai 355B 22T MIT
Released in 355B and 106B sizes. Corpus size was calculated by combining the 15 trillion tokens and the 7 trillion tokens pre-training mix.
GPT-OSS Aug 5 OpenAI 117B Unknown Apache 2.0
Released in 20B and 120B sizes.
Claude 4.1 Aug 5 Anthropic Unknown Unknown Proprietary
Includes one model, Opus.
GPT-5 Aug 7 OpenAI Unknown Unknown
Includes three models: GPT-5, GPT-5 mini, and GPT-5 nano. GPT-5 is available in ChatGPT and API. It includes reasoning abilities.
DeepSeek-V3.1 Aug 21 DeepSeek 671B 15.639T MIT
Corpus size based on DeepSeek V3 (trained on 14.8T tokens); further trained on 839B tokens from the extension phases (630B + 209B). A hybrid model that can switch between thinking and non-thinking modes.
YandexGPT 5.1 Pro Aug 28 Yandex Unknown Unknown Proprietary
Apertus Sep 2 ETH Zurich and EPF Lausanne 70B 15T Apache 2.0
The first LLM to be compliant with the Artificial Intelligence Act of the European Union.
Claude Sonnet 4.5 Sep 29 Anthropic Unknown Unknown Proprietary
GLM-4.6 Sep 30 Z.ai 357B Apache 2.0
Alice AI LLM 1.0 Oct 28 Yandex Unknown Proprietary
Gemini 3 Nov 18 Google DeepMind Unknown
Models released: Deep Think and Pro.
Olmo 3 Nov 20 Allen Institute for AI 32B 5.9T Apache 2.0
Includes 7B and 32B parameter versions, alongside reasoning and instruction-following models.
Claude Opus 4.5 Nov 24 Anthropic Unknown Unknown Proprietary
Largest model in the Claude family.
DeepSeek-V3.2 Dec 1 DeepSeek 685B MIT
Uses a custom DeepSeek Sparse Attention (DSA) mechanism
GPT 5.2 Dec 11 OpenAI Unknown Proprietary
It was able to solve an open problem in statistical learning theory that had previously remained unresolved by human researchers.
GLM-4.7 Dec 22 Z.ai 355B Apache 2.0

2026

Name Release date Developer # of params Corpus size License Notes
Qwen3-Max-Thinking Jan 26 Alibaba Cloud Unknown Unknown Proprietary
Proprietary reasoning model with adaptive tool-use, test-time scaling, and iterative self-reflection.
K2 Think V2 Jan 27 MBUZAI / IFM 70B 12T Apache 2.0
Kimi K2.5 Moonshot AI 1T 15T Modified MIT
Multimodal MoE with 32B active parameters, derived from Kimi K2. Can use "Agent Swarm" technology to coordinate up to 100 parallel sub-agents.
Step 3.5 Flash Feb 12 StepFun 196B Unknown Apache 2.0
MoE model with 11B active parameters out of 196B total
Claude Opus 4.6 Feb 5 Anthropic Unknown Proprietary
GPT-5.3-Codex Feb 5 OpenAI
GLM-5 Feb 12 Z.ai 744B 28.5T MIT
40B active parameters
Qwen3.5 Feb 15 Alibaba Cloud 397B Unknown Apache 2.0
17B active parameters
Qwen3.5-Plus Unknown Proprietary
Claude Sonnet 4.6 Feb 17 Anthropic
Param-2 Feb 17 BharatGen 17B 22T BharatGen Research
Mixture-of-experts model, successor of Param-1; many more Indic languages are supported. Trained on H100 GPUs for 24 days.
Sarvam-105B Feb 18 Sarvam AI 105B 12T Apache 2.0
India's first independently-trained foundation model; has 105B and 30B versions. Based on mixture-of-experts model, using only 10.3B active parameters at a time. Interprets Indic languages and Hinglish. An early checkpoint of the model was released in January.
Sarvam-30B 30B 16T
GPT-5.4 Mar 5 OpenAI Unknown Unknown Proprietary
Mistral Small 4 Mar 17 Mistral AI 119B Apache 2.0
MoE model with 6B active parameters out of 119B total
MiMo-V2-Pro Mar 18 Xiaomi 1T Proprietary
Mixture-of-experts (MoE) model with more than 1 trillion parameters (43 billion active). Designed for agentic scenarios. Initially available on OpenRouter under the codename "Hunter Alpha" before official release.
Qwen3.6-Plus Apr 1 Alibaba Cloud Unknown
Gemma 4 Apr 2 Google DeepMind 31B Apache 2.0
Released in 31B, 26B A4B (3.8B active parameters), E4B (4B effective parameters), and E2B variants
GLM-5.1 Apr 7 Z.ai 744B 28.5T MIT
40B active parameters
Muse Spark Apr 8 Meta Unknown Unknown Proprietary
Qwen3.6 Apr 15 Alibaba Cloud 35B Apache 2.0
MoE model with 3B active parameters out of 35B total
Claude Opus 4.7 Apr 16 Anthropic Unknown Proprietary
Kimi K2.6 Apr 20 Moonshot AI 1T Modified MIT
32B active parameters
GPT-5.5 Apr 23 OpenAI Unknown Proprietary
Hy3 preview Apr 23 Tencent 295B Tencent Hy Community
Mixture-of-experts model with 21B active parameters and a 256K-token context window. Preview predecessor to Hy3.
MiMo-V2.5-Pro Apr 27 Xiaomi 1.02T 48T MIT
MoE model designed for agentic coding and long-horizon software engineering tasks.
MiMo-V2.5 310B 27T
Omni-modal MoE model with agentic capabilities and 1M-token context.
Mistral Medium 3.5 Apr 28 Mistral AI 128B Unknown Modified MIT
Mistral AI's flagship model that merges instruction-following, reasoning, and coding into a single 128B dense model
Composer 2.5 May 18 Cursor 1T Proprietary
Agentic coding model available in the Cursor IDE. Built on Kimi K2.5 with additional reinforcement learning; reported 1T total / 32B active parameters.
Gemini 3.5 Flash May 19 Google DeepMind Unknown
Qwen3.7-Max May 20 Alibaba Cloud
Claude Opus 4.8 May 28 Anthropic
Step 3.7 Flash May 29 StepFun 198B Apache 2.0
Parameter count is 196B (base) + 1.8B (ViT).
MiniMax-M3 Jun 1 MiniMax 428B MiniMax Community
23B active parameters
Nemotron 3 Ultra Jun 4 Nvidia 550B 20T OpenMDW 1.1
55B active parameters
Claude Fable 5 Jun 9 Anthropic Unknown Unknown Proprietary
Kimi K2.7 Code Jun 16 Moonshot AI 1T Modified MIT
32B active parameters
GLM-5.2 Jun 16 Z.ai 744B 28.5T MIT
40B active parameters
LongCat-2.0 Jun 30 Meituan 1.6T 35T
48B active parameters. Preview version previously released as "Owl Alpha" on OpenRouter.
Hy3 Jul 6 Tencent 295B Unknown Apache 2.0
21B active parameters
Grok 4.5 Jul 8 SpaceXAI Unknown Proprietary
Frontier model for coding, agentic tasks, and knowledge work; trained alongside Cursor. Public availability followed on 9 July 2026.
SWE-1.7 Jul 8 Cognition AI
Coding model for the Devin agent; trained from a Kimi K2.7 base with Cognition's reinforcement-learning pipeline. Available via Devin (not as open weights).
Muse Spark 1.1 Jul 9 Meta
GPT-5.6 Jul 9 OpenAI
Inkling Jul 15 Thinking Machines Lab 975B 45T Apache 2.0
41B active parameters
Kimi K3 Jul 16 Moonshot AI 2.8T Unknown Kimi K3
104B active parameters. Weights released on 27 July.
Qwen3.8 Jul 19 Alibaba Cloud 2.4T Various
Available in 2.4T (95B active parameters) and 27B variants. Preview of Qwen3.8-Max (proprietary variant with image input) was released on 19 July. Weights for 2.4T variant released on 12 August. 27B variant released on 14 August.
Gemini 3.5 Flash Lite Jul 21 Google DeepMind Unknown Proprietary
Gemini 3.6 Flash Jul 21 Google DeepMind
Laguna S 2.1 Jul 21 Poolside AI 118B OpenMDW 1.1
8B active parameters
Motif 3 Jul 21 Motif Technologies 314B 12.5T MIT
13.2B active parameters. Preview release on 21 July. Official version and weights released on 13 August.
Solar Open 2 Jul 23 Upstage 250B 12T Upstage Solar
15B active parameters
Fugu-Ultra v1.1 Jul 24 Sakana AI Unknown Unknown Proprietary
Upgraded Fugu-Ultra with stronger coding, agentic, and reasoning benchmarks vs v1.0.
Claude Opus 5 Jul 24 Anthropic
A.X K2 Jul 29 SK Telecom 688B 8.2T Apache 2.0
33B active parameters
Inkling Small Jul 31 Thinking Machines Lab 276B Unknown
12B active parameters
K-Exaone 2.0 Jul 31 LG 750B
37B active parameters
DeepSeek-V4-Flash Jul 31 DeepSeek 284B 32T MIT
13B active parameters. Preview released on 24 April 2026. Official release on 31 July 2026.
Muse Glimmer Aug 10 Meta 30B Unknown Apache 2.0
Gemini 3.7 Flash Aug 13 Google DeepMind Unknown Proprietary
DeepSeek-V4-Pro Aug 13 DeepSeek 1.6T 32T MIT
49B active parameters. Preview released on 24 April 2026. Official release on 13 August 2026.
GLM-5.3 Aug 14 Z.ai 744B 30T Various 40B active
GLM-5.3-flash 28 August 320B 30T MIT 18B active
Hy4 preview Aug 28 Tencent 770B Unknown Apache 2.0
49B active parameters. Preview release.
Claude Mythos 5.1 Sep 1 Anthropic Unknown Proprietary
Claude Fable 5.1 Sep 1 Anthropic
Gemini 3.8 Flash Sep 2 Google DeepMind
GPT-6 Astra Sep 3 OpenAI
K2 Horizon MBUZAI / IFM 0.9B to 375B 20T Apache 2.0
A family of six different models
DeepSeek-V4.1-Flash Sep 10 DeepSeek 552B 45T MIT
8B (prefill) and 16B (decode) active parameters.
Grok 4.7 Sep 21 SpaceXAI Unknown Unknown Proprietary
GPT-6 Sol Sep 22 OpenAI
MiMo-V2.6-Pro Xiaomi 1.02T 30T MIT
Native omni-modal models released as part of the MiMo-V2.6 series.
MiMo-V2.6-Flash 310B 48T
Claude Opus 5.5 Anthropic Unknown Unknown Proprietary
Claude Sonnet 5.5 Sep 28
GPT 6.1 Sep 29 OpenAI
Gemini 4 Argon Sep 30 Google DeepMind
Mistral Large 4 Preview Oct 7 Mistral AI 1.05T Unknown Unknown

See also

Notes

  1. ↑ In many cases, researchers release or report on multiple versions of a model having different sizes. In these cases, the size of the largest model is listed here.
  2. 1 2 3 4 5 6 7 8 9 This is the date that documentation describing the model's architecture was first released.
  3. 1 2 3 4 5 6 7 8 9 License applies only to the pre-trained model weights unless otherwise noted. In many cases, the training code, datasets and post-training data remain proprietary. LLMs may be licensed differently from the chatbots that use them; for the licenses of chatbots, see List of chatbots.
  4. 1 2 3 4 Stack is fully open-source, including training code, datasets and post-training data.
  5. 1 2 Expressed in number of tokens, unless otherwise specified
  6. ↑ Released through the Fairwind Program on 30 September 2026; general launch date was 13 October 2026

References

  1. ↑ "Improving language understanding with unsupervised learning". openai.com. 11 June 2018. Archived from the original on 18 March 2023. Retrieved 18 March 2023.
  2. ↑ "finetune-transformer-lm". GitHub. Archived from the original on 19 May 2023. Retrieved 2 January 2024.
  3. ↑ Radford, Alec (11 June 2018). "Improving language understanding with unsupervised learning". OpenAI. Archived from the original on 7 October 2024. Retrieved 18 November 2025.
  4. 1 2 3 Devlin, Jacob; Chang, Ming-Wei; Lee, Kenton; Toutanova, Kristina (11 October 2018). "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding". arXiv:1810.04805v2 [cs.CL].
  5. ↑ Prickett, Nicole Hemsoth (24 August 2021). "Cerebras Shifts Architecture To Meet Massive AI/ML Models". The Next Platform. Archived from the original on 20 June 2023. Retrieved 20 June 2023.
  6. ↑ "BERT". 13 March 2023. Archived from the original on 13 January 2021. Retrieved 13 March 2023 – via GitHub.
  7. ↑ Manning, Christopher D. (2022). "Human Language Understanding & Reasoning". Daedalus. 151 (2): 127–138. doi:10.1162/daed_a_01905. S2CID 248377870. Archived from the original on 17 November 2023. Retrieved 9 March 2023.
  8. ↑ Patel, Ajay; Li, Bryan; Rasooli, Mohammad Sadegh; Constant, Noah; Raffel, Colin; Callison-Burch, Chris (2022). "Bidirectional Language Models Are Also Few-shot Learners". arXiv:2209.14500 [cs.LG].
  9. 1 2 Raffel, Colin; Shazeer, Noam; Roberts, Adam; Lee, Katherine; Narang, Sharan; Matena, Michael; Zhou, Yanqi; Li, Wei; Liu, Peter J. (2020). "Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer". Journal of Machine Learning Research. 21 (140): 1–67. arXiv:1910.10683. ISSN 1533-7928. Archived from the original on 5 October 2024. Retrieved 26 December 2024.
  10. ↑ google-research/text-to-text-transfer-transformer, Google Research, 2 April 2024, archived from the original on 29 March 2024, retrieved 4 April 2024
  11. ↑ "Imagen: Text-to-Image Diffusion Models". imagen.research.google. Archived from the original on 27 March 2024. Retrieved 4 April 2024.
  12. ↑ "Pretrained models — transformers 2.0.0 documentation". huggingface.co. Archived from the original on 5 August 2024. Retrieved 5 August 2024.
  13. ↑ "xlnet". GitHub. Archived from the original on 2 January 2024. Retrieved 2 January 2024.
  14. ↑ Yang, Zhilin; Dai, Zihang; Yang, Yiming; Carbonell, Jaime; Salakhutdinov, Ruslan; Le, Quoc V. (2 January 2020). "XLNet: Generalized Autoregressive Pretraining for Language Understanding". arXiv:1906.08237 [cs.CL].
  15. ↑ "GPT-2: 1.5B Release". OpenAI. 5 November 2019. Archived from the original on 14 November 2019. Retrieved 14 November 2019.
  16. ↑ "Better language models and their implications". openai.com. Archived from the original on 16 March 2023. Retrieved 13 March 2023.
  17. 1 2 "OpenAI's GPT-3 Language Model: A Technical Overview". lambdalabs.com. 3 June 2020. Archived from the original on 27 March 2023. Retrieved 13 March 2023.
  18. 1 2 "openai-community/gpt2-xl · Hugging Face". huggingface.co. 21 October 2019. Archived from the original on 24 July 2024. Retrieved 24 July 2024.
  19. ↑ "gpt-2". GitHub. Archived from the original on 11 March 2023. Retrieved 13 March 2023.
  20. ↑ Wiggers, Kyle (28 April 2022). "The emerging types of language models and why they matter". TechCrunch. Archived from the original on 16 March 2023. Retrieved 9 March 2023.
  21. ↑ Table D.1 in Brown, Tom B.; Mann, Benjamin; Ryder, Nick; Subbiah, Melanie; Kaplan, Jared; Dhariwal, Prafulla; Neelakantan, Arvind; Shyam, Pranav; Sastry, Girish; Askell, Amanda; Agarwal, Sandhini; Herbert-Voss, Ariel; Krueger, Gretchen; Henighan, Tom; Child, Rewon; Ramesh, Aditya; Ziegler, Daniel M.; Wu, Jeffrey; Winter, Clemens; Hesse, Christopher; Chen, Mark; Sigler, Eric; Litwin, Mateusz; Gray, Scott; Chess, Benjamin; Clark, Jack; Berner, Christopher; McCandlish, Sam; Radford, Alec; Sutskever, Ilya; Amodei, Dario (28 May 2020). "Language Models are Few-Shot Learners". arXiv:2005.14165v4 [cs.CL].
  22. ↑ "ChatGPT: Optimizing Language Models for Dialogue". OpenAI. 30 November 2022. Archived from the original on 30 November 2022. Retrieved 13 January 2023.
  23. ↑ "GPT Neo". 15 March 2023. Archived from the original on 12 March 2023. Retrieved 12 March 2023 – via GitHub.
  24. 1 2 3 Gao, Leo; Biderman, Stella; Black, Sid; Golding, Laurence; Hoppe, Travis; Foster, Charles; Phang, Jason; He, Horace; Thite, Anish; Nabeshima, Noa; Presser, Shawn; Leahy, Connor (31 December 2020). "The Pile: An 800GB Dataset of Diverse Text for Language Modeling". arXiv:2101.00027 [cs.CL].
  25. 1 2 Iyer, Abhishek (15 May 2021). "GPT-3's free alternative GPT-Neo is something to be excited about". VentureBeat. Archived from the original on 9 March 2023. Retrieved 13 March 2023.
  26. ↑ Kim, Han-joo (25 May 2021). "Naver unveils supersized AI platform HyperCLOVA". Yonhap News Agency. Archived from the original on 15 June 2026. Retrieved 15 June 2026.
  27. 1 2 Wiggers, Kyle (1 June 2021). "Naver trained a 'GPT-3-like' Korean language model". VentureBeat ∣ Transformative tech coverage that matters. Archived from the original on 20 September 2022. Retrieved 15 June 2026.
  28. 1 2 Kim, Boseop; Kim, HyoungSeok; Lee, Sang-Woo; Lee, Gichang; Kwak, Donghyun; Dong Hyeon, Jeon; Park, Sunghyun; Kim, Sungju; Kim, Seonhoon; Seo, Dongpil; Lee, Heungsub; Jeong, Minyoung; Lee, Sungjae; Kim, Minsub; Ko, Suk Hyun; Kim, Seokhun; Park, Taeyong; Kim, Jinuk; Kang, Soyoung; Ryu, Na-Hyeon; Yoo, Kang Min; Chang, Minsuk; Suh, Soobin; In, Sookyo; Park, Jinseong; Kim, Kyungduk; Kim, Hiun; Jeong, Jisu; Yeo, Yong Goo; Ham, Donghoon; Park, Dongju; Lee, Min Young; Kang, Jaewook; Kang, Inho; Ha, Jung-Woo; Park, Woomyoung; Sung, Nako (November 2021). "What Changes Can Large-scale Language Models Bring? Intensive Study on HyperCLOVA: Billions-scale Korean Generative Pretrained Transformers". In Moens, Marie-Francine; Huang, Xuanjing; Specia, Lucia; Yih, Scott Wen-tau (eds.). Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. Online and Punta Cana, Dominican Republic: Association for Computational Linguistics. pp. 3405–3424. arXiv:2109.04650. doi:10.18653/v1/2021.emnlp-main.274.
  29. ↑ "GPT-J-6B: An Introduction to the Largest Open Source GPT Model | Forefront". www.forefront.ai. Archived from the original on 9 March 2023. Retrieved 28 February 2023.
  30. 1 2 3 4 Dey, Nolan; Gosal, Gurpreet; Zhiming; Chen; Khachane, Hemant; Marshall, William; Pathria, Ribhu; Tom, Marvin; Hestness, Joel (1 April 2023). "Cerebras-GPT: Open Compute-Optimal Language Models Trained on the Cerebras Wafer-Scale Cluster". arXiv:2304.03208 [cs.LG].
  31. ↑ Alvi, Ali; Kharya, Paresh (11 October 2021). "Using DeepSpeed and Megatron to Train Megatron-Turing NLG 530B, the World's Largest and Most Powerful Generative Language Model". Microsoft Research. Archived from the original on 13 March 2023. Retrieved 13 March 2023.
  32. 1 2 Smith, Shaden; Patwary, Mostofa; Norick, Brandon; LeGresley, Patrick; Rajbhandari, Samyam; Casper, Jared; Liu, Zhun; Prabhumoye, Shrimai; Zerveas, George; Korthikanti, Vijay; Zhang, Elton; Child, Rewon; Aminabadi, Reza Yazdani; Bernauer, Julie; Song, Xia (4 February 2022). "Using DeepSpeed and Megatron to Train Megatron-Turing NLG 530B, A Large-Scale Generative Language Model". arXiv:2201.11990 [cs.CL].
  33. 1 2 Rajbhandari, Samyam; Li, Conglong; Yao, Zhewei; Zhang, Minjia; Aminabadi, Reza Yazdani; Awan, Ammar Ahmad; Rasley, Jeff; He, Yuxiong (21 July 2022), DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and Training to Power Next-Generation AI Scale, arXiv:2201.05596
  34. ↑ Wang, Shuohuan; Sun, Yu; Xiang, Yang; Wu, Zhihua; Ding, Siyu; Gong, Weibao; Feng, Shikun; Shang, Junyuan; Zhao, Yanbin; Pang, Chao; Liu, Jiaxiang; Chen, Xuyi; Lu, Yuxiang; Liu, Weixin; Wang, Xi; Bai, Yangfan; Chen, Qiuliang; Zhao, Li; Li, Shiyong; Sun, Peng; Yu, Dianhai; Ma, Yanjun; Tian, Hao; Wu, Hua; Wu, Tian; Zeng, Wei; Li, Ge; Gao, Wen; Wang, Haifeng (23 December 2021). "ERNIE 3.0 Titan: Exploring Larger-scale Knowledge Enhanced Pre-training for Language Understanding and Generation". arXiv:2112.12731 [cs.CL].
  35. ↑ "Product". Anthropic. Archived from the original on 16 March 2023. Retrieved 14 March 2023.
  36. 1 2 Askell, Amanda; Bai, Yuntao; Chen, Anna; et al. (9 December 2021). "A General Language Assistant as a Laboratory for Alignment". arXiv:2112.00861 [cs.CL].
  37. ↑ Bai, Yuntao; Kadavath, Saurav; Kundu, Sandipan; et al. (15 December 2022). "Constitutional AI: Harmlessness from AI Feedback". arXiv:2212.08073 [cs.CL].
  38. 1 2 3 Dai, Andrew M; Du, Nan (9 December 2021). "More Efficient In-Context Learning with GLaM". ai.googleblog.com. Archived from the original on 12 March 2023. Retrieved 9 March 2023.
  39. ↑ "Language modelling at scale: Gopher, ethical considerations, and retrieval". www.deepmind.com. 8 December 2021. Archived from the original on 20 March 2023. Retrieved 20 March 2023.
  40. 1 2 3 Hoffmann, Jordan; Borgeaud, Sebastian; Mensch, Arthur; et al. (29 March 2022). "Training Compute-Optimal Large Language Models". arXiv:2203.15556 [cs.CL].
  41. 1 2 3 4 Table 20 and page 66 of PaLM: Scaling Language Modeling with Pathways Archived 10 June 2023 at the Wayback Machine
  42. 1 2 Cheng, Heng-Tze; Thoppilan, Romal (21 January 2022). "LaMDA: Towards Safe, Grounded, and High-Quality Dialog Models for Everything". ai.googleblog.com. Archived from the original on 25 March 2022. Retrieved 9 March 2023.
  43. ↑ Thoppilan, Romal; De Freitas, Daniel; Hall, Jamie; Shazeer, Noam; Kulshreshtha, Apoorv; Cheng, Heng-Tze; Jin, Alicia; Bos, Taylor; Baker, Leslie; Du, Yu; Li, YaGuang; Lee, Hongrae; Zheng, Huaixiu Steven; Ghafouri, Amin; Menegali, Marcelo (1 January 2022). "LaMDA: Language Models for Dialog Applications". arXiv:2201.08239 [cs.CL].
  44. ↑ Black, Sidney; Biderman, Stella; Hallahan, Eric; et al. (1 May 2022). GPT-NeoX-20B: An Open-Source Autoregressive Language Model. Proceedings of BigScience Episode #5 – Workshop on Challenges & Perspectives in Creating Large Language Models. Vol. Proceedings of BigScience Episode #5 – Workshop on Challenges & Perspectives in Creating Large Language Models. pp. 95–136. Archived from the original on 10 December 2022. Retrieved 19 December 2022.
  45. 1 2 3 Hoffmann, Jordan; Borgeaud, Sebastian; Mensch, Arthur; Sifre, Laurent (12 April 2022). "An empirical analysis of compute-optimal large language model training". Deepmind Blog. Archived from the original on 13 April 2022. Retrieved 9 March 2023.
  46. ↑ Narang, Sharan; Chowdhery, Aakanksha (4 April 2022). "Pathways Language Model (PaLM): Scaling to 540 Billion Parameters for Breakthrough Performance". ai.googleblog.com. Archived from the original on 4 April 2022. Retrieved 9 March 2023.
  47. ↑ Zhang, Susan; Diab, Mona; Zettlemoyer, Luke. "Democratizing access to large-scale language models with OPT-175B". ai.facebook.com. Archived from the original on 12 March 2023. Retrieved 12 March 2023.
  48. ↑ Zhang, Susan; Roller, Stephen; Goyal, Naman; Artetxe, Mikel; Chen, Moya; Chen, Shuohui; Dewan, Christopher; Diab, Mona; Li, Xian; Lin, Xi Victoria; Mihaylov, Todor; Ott, Myle; Shleifer, Sam; Shuster, Kurt; Simig, Daniel; Koura, Punit Singh; Sridhar, Anjali; Wang, Tianlu; Zettlemoyer, Luke (21 June 2022). "OPT: Open Pre-trained Transformer Language Models". arXiv:2205.01068 [cs.CL].
  49. ↑ "metaseq/projects/OPT/chronicles at main · facebookresearch/metaseq". GitHub. Archived from the original on 24 January 2024. Retrieved 18 October 2024.
  50. 1 2 Khrushchev, Mikhail; Vasilev, Ruslan; Petrov, Alexey; Zinov, Nikolay (22 June 2022), YaLM 100B, archived from the original on 16 June 2023, retrieved 18 March 2023
  51. 1 2 Lewkowycz, Aitor; Andreassen, Anders; Dohan, David; Dyer, Ethan; Michalewski, Henryk; Ramasesh, Vinay; Slone, Ambrose; Anil, Cem; Schlag, Imanol; Gutman-Solo, Theo; Wu, Yuhuai; Neyshabur, Behnam; Gur-Ari, Guy; Misra, Vedant (30 June 2022). "Solving Quantitative Reasoning Problems with Language Models". arXiv:2206.14858 [cs.CL].
  52. ↑ "Minerva: Solving Quantitative Reasoning Problems with Language Models". ai.googleblog.com. 30 June 2022. Archived from the original on 20 March 2023. Retrieved 20 March 2023.
  53. ↑ Ananthaswamy, Anil (8 March 2023). "In AI, is bigger always better?". Nature. 615 (7951): 202–205. Bibcode:2023Natur.615..202A. doi:10.1038/d41586-023-00641-w. PMID 36890378. S2CID 257380916. Archived from the original on 16 March 2023. Retrieved 9 March 2023.
  54. ↑ "bigscience/bloom · Hugging Face". huggingface.co. 9 November 2022. Archived from the original on 12 April 2023. Retrieved 13 March 2023.
  55. ↑ Taylor, Ross; Kardas, Marcin; Cucurull, Guillem; Scialom, Thomas; Hartshorn, Anthony; Saravia, Elvis; Poulton, Andrew; Kerkez, Viktor; Stojnic, Robert (16 November 2022). "Galactica: A Large Language Model for Science". arXiv:2211.09085 [cs.CL].
  56. ↑ "20B-parameter Alexa model sets new marks in few-shot learning". Amazon Science. 2 August 2022. Archived from the original on 15 March 2023. Retrieved 12 March 2023.
  57. ↑ Soltan, Saleh; Ananthakrishnan, Shankar; FitzGerald, Jack; et al. (3 August 2022). "AlexaTM 20B: Few-Shot Learning Using a Large-Scale Multilingual Seq2Seq Model". arXiv:2208.01448 [cs.CL].
  58. ↑ "AlexaTM 20B is now available in Amazon SageMaker JumpStart | AWS Machine Learning Blog". aws.amazon.com. 17 November 2022. Archived from the original on 13 March 2023. Retrieved 13 March 2023.
  59. 1 2 "Introducing LLaMA: A foundational, 65-billion-parameter large language model". Meta AI. 24 February 2023. Archived from the original on 3 March 2023. Retrieved 9 March 2023.
  60. 1 2 3 "The Falcon has landed in the Hugging Face ecosystem". huggingface.co. 19 May 2023. Archived from the original on 20 June 2023. Retrieved 20 June 2023.
  61. ↑ Taori, Rohan; Gulrajani, Ishaan; Zhang, Tianyi; Dubois, Yann; Li, Xuechen; Guestrin, Carlos; Liang, Percy; Hashimoto, Tatsunori B. (13 March 2023). "Alpaca: A Strong, Replicable Instruction-Following Model". Stanford Center for Research on Foundation Models. Stanford University. Archived from the original on 6 April 2023. Retrieved 13 September 2026.
  62. ↑ Dey, Nolan (28 March 2023). "Cerebras-GPT: A Family of Open, Compute-efficient, Large Language Models". Cerebras. Archived from the original on 28 March 2023. Retrieved 28 March 2023.
  63. ↑ "Abu Dhabi-based TII launches its own version of ChatGPT". tii.ae. Archived from the original on 3 April 2023. Retrieved 3 April 2023.
  64. ↑ Penedo, Guilherme; Malartic, Quentin; Hesslow, Daniel; Cojocaru, Ruxandra; Cappelli, Alessandro; Alobeidli, Hamza; Pannier, Baptiste; Almazrouei, Ebtesam; Launay, Julien (1 June 2023). "The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora with Web Data, and Web Data Only". arXiv:2306.01116 [cs.CL].
  65. ↑ "tiiuae/falcon-40b · Hugging Face". huggingface.co. 9 June 2023. Archived from the original on 20 June 2023. Retrieved 20 June 2023.
  66. ↑ UAE's Falcon 40B, World's Top-Ranked AI Model from Technology Innovation Institute, is Now Royalty-Free Archived 8 February 2024 at the Wayback Machine, 31 May 2023
  67. 1 2 Wu, Shijie; Irsoy, Ozan; Lu, Steven; Dabravolski, Vadim; Dredze, Mark; Gehrmann, Sebastian; Kambadur, Prabhanjan; Rosenberg, David; Mann, Gideon (30 March 2023). "BloombergGPT: A Large Language Model for Finance". arXiv:2303.17564 [cs.LG].
  68. ↑ Ren, Xiaozhe; Zhou, Pingyi; Meng, Xinfan; Huang, Xinjing; Wang, Yadao; Wang, Weichao; Li, Pengfei; Zhang, Xiaoda; Podolskiy, Alexander; Arshinov, Grigory; Bout, Andrey; Piontkovskaya, Irina; Wei, Jiansheng; Jiang, Xin; Su, Teng; Liu, Qun; Yao, Jun (19 March 2023). "PanGu-Σ: Towards Trillion Parameter Language Model with Sparse Heterogeneous Computing". arXiv:2303.10845 [cs.CL].
  69. ↑ Köpf, Andreas; Kilcher, Yannic; von Rütte, Dimitri; Anagnostidis, Sotiris; Tam, Zhi-Rui; Stevens, Keith; Barhoum, Abdullah; Duc, Nguyen Minh; Stanley, Oliver; Nagyfi, Richárd; ES, Shahul; Suri, Sameer; Glushkov, David; Dantuluri, Arnav; Maguire, Andrew (14 April 2023). "OpenAssistant Conversations – Democratizing Large Language Model Alignment". arXiv:2304.07327 [cs.CL].
  70. ↑ Chiang, Wei-Lin; Li, Zhuohan; Lin, Zi; Sheng, Ying; Wu, Zhanghao; Zhang, Hao; Zheng, Lianmin; Zhuang, Siyuan; Zhuang, Yonghao; Gonzalez, Joseph E.; Stoica, Ion; Xing, Eric P. (30 March 2023). "Vicuna: An Open-Source Chatbot Impressing GPT-4 with 90%* ChatGPT Quality". LMSYS Org. Retrieved 13 September 2026.
  71. ↑ Wrobel, Sharon. "Tel Aviv startup rolls out new advanced AI language model to rival OpenAI". The Times of Israel. ISSN 0040-7909. Archived from the original on 24 July 2023. Retrieved 24 July 2023.
  72. ↑ Wiggers, Kyle (13 April 2023). "With Bedrock, Amazon enters the generative AI race". TechCrunch. Archived from the original on 24 July 2023. Retrieved 24 July 2023.
  73. 1 2 Elias, Jennifer (16 May 2023). "Google's newest A.I. model uses nearly five times more text data for training than its predecessor". CNBC. Archived from the original on 16 May 2023. Retrieved 18 May 2023.
  74. ↑ "Introducing PaLM 2". Google. 10 May 2023. Archived from the original on 18 May 2023. Retrieved 18 May 2023.
  75. 1 2 3 Gunasekar, Suriya; Zhang, Yi; Aneja, Jyoti; Caio César Teodoro Mendes; Allie Del Giorno; Gopi, Sivakanth; Javaheripi, Mojan; Kauffmann, Piero; Gustavo de Rosa; Saarikivi, Olli; Salim, Adil; Shah, Shital; Harkirat Singh Behl; Wang, Xin; Bubeck, Sébastien; Eldan, Ronen; Adam Tauman Kalai; Yin Tat Lee; Li, Yuanzhi (2023). "Textbooks Are All You Need". arXiv:2306.11644 [cs.CL].
  76. 1 2 "Introducing Llama 2: The Next Generation of Our Open Source Large Language Model". Meta AI. 2023. Archived from the original on 5 January 2024. Retrieved 19 July 2023.
  77. ↑ "llama/MODEL_CARD.md at main · meta-llama/llama". GitHub. Archived from the original on 28 May 2024. Retrieved 28 May 2024.
  78. ↑ "Claude 2". anthropic.com. Archived from the original on 15 December 2023. Retrieved 12 December 2023.
  79. ↑ Nirmal, Dinesh (7 September 2023). "Building AI for business: IBM's Granite foundation models". IBM Blog. Archived from the original on 22 July 2024. Retrieved 11 August 2024.
  80. ↑ "Announcing Mistral 7B". Mistral. 2023. Archived from the original on 6 January 2024. Retrieved 6 October 2023.
  81. ↑ "Introducing Claude 2.1". anthropic.com. Archived from the original on 15 December 2023. Retrieved 12 December 2023.
  82. ↑ xai-org/grok-1, xai-org, 19 March 2024, archived from the original on 28 May 2024, retrieved 19 March 2024
  83. ↑ "Grok-1 model card". x.ai. Archived from the original on 11 December 2023. Retrieved 12 December 2023.
  84. ↑ "Gemini – Google DeepMind". deepmind.google. Archived from the original on 8 December 2023. Retrieved 12 December 2023.
  85. ↑ Franzen, Carl (11 December 2023). "Mistral shocks AI community as latest open source model eclipses GPT-3.5 performance". VentureBeat. Archived from the original on 11 December 2023. Retrieved 12 December 2023.
  86. ↑ "Mixtral of experts". mistral.ai. 11 December 2023. Archived from the original on 13 February 2024. Retrieved 12 December 2023.
  87. 1 2 DeepSeek-AI; Bi, Xiao; Chen, Deli; Chen, Guanting; Chen, Shanhuang; Dai, Damai; Deng, Chengqi; Ding, Honghui; Dong, Kai (5 January 2024). "DeepSeek LLM: Scaling Open-Source Language Models with Longtermism". arXiv:2401.02954 [cs.CL].
  88. 1 2 Hughes, Alyssa (12 December 2023). "Phi-2: The surprising power of small language models". Microsoft Research. Archived from the original on 12 December 2023. Retrieved 13 December 2023.
  89. ↑ "Our next-generation model: Gemini 1.5". Google. 15 February 2024. Archived from the original on 16 February 2024. Retrieved 16 February 2024. This means 1.5 Pro can process vast amounts of information in one go — including 1 hour of video, 11 hours of audio, codebases with over 30,000 lines of code or over 700,000 words. In our research, we've also successfully tested up to 10 million tokens.
  90. ↑ "Gemma". Archived from the original on 5 January 2025. Retrieved 26 December 2024 – via GitHub.
  91. ↑ "OLMo: Open Language Model | Ai2". allenai.org. Archived from the original on 13 March 2026. Retrieved 17 March 2026.
  92. ↑ Groeneveld, Dirk; Beltagy, Iz; Walsh, Pete; Bhagia, Akshita; Kinney, Rodney; Tafjord, Oyvind; Jha, Ananya Harsh; Ivison, Hamish; Magnusson, Ian (7 June 2024). "OLMo: Accelerating the Science of Language Models". arXiv:2402.00838 [cs.CL].
  93. ↑ "Introducing the next generation of Claude". Anthropic. Archived from the original on 4 March 2024. Retrieved 4 March 2024.
  94. ↑ "Databricks Open Model License". Databricks. 27 March 2024. Archived from the original on 25 August 2025. Retrieved 6 August 2025.
  95. ↑ "Databricks Open Model Acceptable Use Policy". Databricks. 27 March 2024. Archived from the original on 16 September 2025. Retrieved 6 August 2025.
  96. 1 2 "Release of "Fugaku-LLM" - a large language model trained on the supercomputer "Fugaku"". Fujitsu. 10 May 2024. Archived from the original on 19 May 2026. Retrieved 20 April 2026.
  97. ↑ "Fugaku-LLM Terms of Use". 23 April 2024. Archived from the original on 7 September 2025. Retrieved 6 August 2025 – via Hugging Face.
  98. ↑ "Fugaku-LLM/Fugaku-LLM-13B · Hugging Face". huggingface.co. Archived from the original on 17 May 2024. Retrieved 17 May 2024.
  99. ↑ Dickson, Ben (22 May 2024). "Meta introduces Chameleon, a state-of-the-art multimodal model". VentureBeat. Archived from the original on 11 February 2025. Retrieved 26 December 2024.
  100. ↑ "chameleon/LICENSE at e3b711ef63b0bb3a129cf0cf0918e36a32f26e2c · facebookresearch/chameleon". Meta Research. Archived from the original on 15 September 2025. Retrieved 6 August 2025 – via GitHub.
  101. ↑ AI, Mistral (17 April 2024). "Cheaper, Better, Faster, Stronger". mistral.ai. Archived from the original on 5 May 2024. Retrieved 5 May 2024.
  102. 1 2 Bilenko, Misha (23 April 2024). "Introducing Phi-3: Redefining what's possible with SLMs". azure.microsoft.com. Archived from the original on 8 May 2026. Retrieved 8 May 2026.
  103. ↑ Abdin, Marah; et al. (2024). "Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone". arXiv:2404.14219 [cs.CL].
  104. ↑ "Qwen2". GitHub. Archived from the original on 17 June 2024. Retrieved 17 June 2024.
  105. ↑ DeepSeek-AI; Liu, Aixin; Feng, Bei; Wang, Bin; Wang, Bingxuan; Liu, Bo; Zhao, Chenggang; Dengr, Chengqi; Ruan, Chong (19 June 2024). "DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model". arXiv:2405.04434 [cs.CL].
  106. ↑ "NVIDIA Open Models License". Nvidia. 16 June 2025. Retrieved 6 August 2025.
  107. ↑ "Trustworthy AI". Nvidia. 27 June 2024. Archived from the original on 3 August 2025. Retrieved 6 August 2025.
  108. ↑ "nvidia/Nemotron-4-340B-Base · Hugging Face". huggingface.co. 14 June 2024. Archived from the original on 15 June 2024. Retrieved 15 June 2024.
  109. ↑ "Nemotron-4 340B | Research". research.nvidia.com. 14 June 2024. Archived from the original on 15 June 2024. Retrieved 15 June 2024.
  110. ↑ "Introducing Claude 3.5 Sonnet". Anthropic. Archived from the original on 15 July 2026. Retrieved 8 August 2025.
  111. ↑ "Introducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku". Anthropic. Archived from the original on 15 July 2026. Retrieved 8 August 2025.
  112. ↑ "The Llama 3 Herd of Models" (July 23, 2024) Llama Team, AI @ Meta
  113. ↑ "llama-models/models/llama3_1/MODEL_CARD.md at main · meta-llama/llama-models". GitHub. Archived from the original on 23 July 2024. Retrieved 23 July 2024.
  114. ↑ "LICENSE · xai-org/grok-2 at main". 5 November 2025. Archived from the original on 5 December 2025. Retrieved 18 November 2025 – via Hugging Face.
  115. ↑ "xAI Acceptable Use Policy". xAI. 2 January 2025. Retrieved 18 November 2025.
  116. ↑ Weatherbed, Jess (14 August 2024). "xAI's new Grok-2 chatbots bring AI image generation to X". The Verge. Archived from the original on 17 November 2024. Retrieved 18 November 2025.
  117. ↑ Ha, Anthony (24 August 2025). "Elon Musk says xAI has open sourced Grok 2.5". TechCrunch. Retrieved 18 November 2025.
  118. ↑ "Introducing OpenAI o1". openai.com. Archived from the original on 8 February 2025. Retrieved 8 August 2025.
  119. ↑ Paul, Katie; Tong, Anna (13 September 2024). "OpenAI launches new series of AI models with 'reasoning' abilities". Reuters.
  120. ↑ Jindal, Siddharth (24 October 2024). "Sarvam AI Launches Sarvam-1, Outperforms Gemma-2 and Llama-3.2". Analytics India Magazine. Archived from the original on 25 July 2025. Retrieved 20 April 2026.
  121. ↑ "LICENSE.md · sarvamai/sarvam-1". 23 October 2024. Retrieved 20 April 2026 – via Hugging Face.
  122. 1 2 "Models Overview". mistral.ai. Archived from the original on 9 April 2025. Retrieved 3 March 2025.
  123. ↑ "OLMo 2: The best fully open language model to date | Ai2". allenai.org. Archived from the original on 11 March 2026. Retrieved 17 March 2026.
  124. 1 2 3 OLMo, Team; Walsh, Pete; Soldaini, Luca; Groeneveld, Dirk; Lo, Kyle; Arora, Shane; Bhagia, Akshita; Gu, Yuling; Huang, Shengyi (8 October 2025). "2 OLMo 2 Furious". arXiv:2501.00656 [cs.CL].
  125. ↑ "Phi-4 Model Card". huggingface.co. 10 July 2025. Archived from the original on 18 November 2025. Retrieved 11 November 2025.
  126. ↑ "Introducing Phi-4: Microsoft's Newest Small Language Model Specializing in Complex Reasoning". techcommunity.microsoft.com. Archived from the original on 13 November 2025. Retrieved 11 November 2025.
  127. ↑ deepseek-ai/DeepSeek-V3, DeepSeek, 26 December 2024, archived from the original on 27 March 2025, retrieved 26 December 2024
  128. ↑ Feng, Coco (25 March 2025). "DeepSeek wows coders with more powerful open-source V3 model". South China Morning Post. Retrieved 6 April 2025.
  129. ↑ Amazon Nova Micro, Lite, and Pro - AWS AI Service Cards3, Amazon, 27 December 2024, archived from the original on 11 February 2025, retrieved 27 December 2024
  130. ↑ deepseek-ai/DeepSeek-R1, DeepSeek, 21 January 2025, archived from the original on 20 January 2025, retrieved 21 January 2025
  131. ↑ DeepSeek-AI; Guo, Daya; Yang, Dejian; Zhang, Haowei; Song, Junxiao; Zhang, Ruoyu; Xu, Runxin; Zhu, Qihao; Ma, Shirong (22 January 2025), "DeepSeek-R1 incentivizes reasoning in LLMS through reinforcement learning", Nature, 645 (8081): 633–638, arXiv:2501.12948, Bibcode:2025Natur.645..633G, doi:10.1038/s41586-025-09422-z, PMC 12443585, PMID 40962978
  132. ↑ Qwen; Yang, An; Yang, Baosong; Zhang, Beichen; Hui, Binyuan; Zheng, Bo; Yu, Bowen; Li, Chengyuan; Liu, Dayiheng (3 January 2025). "Qwen2.5 Technical Report". arXiv:2412.15115 [cs.CL].
  133. ↑ MiniMax-AI/MiniMax-01, MiniMax, 26 January 2025, archived from the original on 28 January 2025, retrieved 26 January 2025
  134. ↑ MiniMax; Li, Aonian; Gong, Bangwei; Yang, Bo; Shan, Boji; Liu, Chang; Zhu, Cheng; Zhang, Chunhao; Guo, Congchao (14 January 2025). "MiniMax-01: Scaling Foundation Models with Lightning Attention". arXiv:2501.08313 [cs.CL].
  135. ↑ Kavukcuoglu, Koray (5 February 2025). "Gemini 2.0 is now available to everyone". Google. Archived from the original on 10 April 2025. Retrieved 6 February 2025.
  136. ↑ "Gemini 2.0: Flash, Flash-Lite and Pro". Google for Developers. Archived from the original on 10 April 2025. Retrieved 6 February 2025.
  137. ↑ Franzen, Carl (5 February 2025). "Google launches Gemini 2.0 Pro, Flash-Lite and connects reasoning model Flash Thinking to YouTube, Maps and Search". VentureBeat. Archived from the original on 17 March 2025. Retrieved 6 February 2025.
  138. ↑ "Grok 3 Beta — The Age of Reasoning Agents". x.ai. Archived from the original on 22 February 2025. Retrieved 22 February 2025.
  139. ↑ "Claude 3.7 Sonnet and Claude Code". Anthropic. Archived from the original on 24 February 2025. Retrieved 8 August 2025.
  140. ↑ "Introducing GPT-4.5". openai.com. Archived from the original on 27 February 2025. Retrieved 8 August 2025.
  141. ↑ Kavukcuoglu, Koray (25 March 2025). "Gemini 2.5: Our most intelligent AI model". Google. Archived from the original on 5 January 2026. Retrieved 23 September 2025.
  142. ↑ "meta-llama/Llama-4-Maverick-17B-128E · Hugging Face". huggingface.co. 5 April 2025. Archived from the original on 21 May 2026. Retrieved 6 April 2025.
  143. ↑ "The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation". ai.meta.com. Archived from the original on 5 April 2025. Retrieved 5 April 2025.
  144. ↑ "Introducing OpenAI o3 and o4-mini". openai.com. Archived from the original on 16 April 2025. Retrieved 8 August 2025.
  145. ↑ Team, Qwen (29 April 2025). "Qwen3: Think Deeper, Act Faster". Qwen. Archived from the original on 29 April 2025. Retrieved 29 April 2025.
  146. ↑ "Introducing Claude 4". Anthropic. Archived from the original on 5 August 2025. Retrieved 8 August 2025.
  147. ↑ Yadav, Nandini (26 May 2025). "Indian AI startup launches Sarvam-M model: What is it, why is everyone talking about it". India Today. Archived from the original on 6 October 2025. Retrieved 18 March 2026.
  148. ↑ "Sarvam-M: Open Source Hybrid Indic LLM | Sarvam AI". Sarvam AI. 23 May 2025. Retrieved 18 March 2026.
  149. ↑ "Grok 4". x.ai. 9 July 2025. Archived from the original on 9 May 2026. Retrieved 9 May 2026.
  150. ↑ Pundalik, Kundeshwar; Sawarkar, Piyush; Sahoo, Nihar; Shinde, Abhishek; Chanda, Prateek; Goswami, Vedant; Nagpal, Ajay; Singh, Atul; Thakur, Viraj (16 July 2025). "PARAM-1 BharatGen 2.9B Model". arXiv:2507.13390 [cs.CL].
  151. ↑ "README.md · bharatgenai/Param-1". 24 February 2026. Archived from the original on 22 April 2026. Retrieved 12 April 2026 – via Hugging Face.
  152. ↑ "zai-org/GLM-4.5 · Hugging Face". huggingface.co. 4 August 2025. Archived from the original on 6 August 2025. Retrieved 6 August 2025.
  153. ↑ "GLM-4.5: Reasoning, Coding, and Agentic Abililties". z.ai. Retrieved 6 August 2025.
  154. ↑ Whitwam, Ryan (5 August 2025). "OpenAI announces two "gpt-oss" open AI models, and you can download them today". Ars Technica. Archived from the original on 6 August 2025. Retrieved 6 August 2025.
  155. ↑ "Claude Opus 4.1". Anthropic. Archived from the original on 15 July 2026. Retrieved 8 August 2025.
  156. ↑ "Introducing GPT-5". openai.com. 7 August 2025. Archived from the original on 9 November 2025. Retrieved 8 August 2025.
  157. ↑ "OpenAI Platform: GPT-5 Model Documentation". openai.com. Archived from the original on 17 August 2025. Retrieved 18 August 2025.
  158. ↑ "deepseek-ai/DeepSeek-V3.1 · Hugging Face". huggingface.co. 21 August 2025. Archived from the original on 25 August 2025. Retrieved 25 August 2025.
  159. ↑ "DeepSeek-V3.1 Release | DeepSeek API Docs". api-docs.deepseek.com. Archived from the original on 24 August 2025. Retrieved 25 August 2025.
  160. ↑ Kirchner, Malte (2 September 2025). "Apertus: Schweiz stellt erstes offenes und mehrsprachiges KI-Modell vor". heise online (in German). Retrieved 7 November 2025.
  161. ↑ "Apertus: Ein vollständig offenes, transparentes und mehrsprachiges Sprachmodell" (in German). Zürich: ETH Zürich. 2 September 2025. Retrieved 7 November 2025.
  162. ↑ "Introducing Claude Sonnet 4.5". Anthropic. Archived from the original on 17 July 2026. Retrieved 29 September 2025.
  163. ↑ "GLM-4.6: Advanced Agentic, Reasoning and Coding Capabilities". z.ai. Retrieved 1 October 2025.
  164. ↑ "zai-org/GLM-4.6 · Hugging Face". huggingface.co. 30 September 2025. Archived from the original on 30 September 2025. Retrieved 1 October 2025.
  165. ↑ "GLM-4.6". modelscope.cn. Retrieved 1 October 2025.
  166. ↑ "A new era of intelligence with Gemini 3". Google. 18 November 2025. Archived from the original on 29 November 2025. Retrieved 5 January 2026.
  167. ↑ "Olmo 3: Charting a path through the model flow to lead open-source AI". Ai2. 20 November 2025. Archived from the original on 15 March 2026. Retrieved 17 March 2026.
  168. ↑ Olmo, Team; Ettinger, Allyson; Bertsch, Amanda; Kuehl, Bailey; Graham, David; Heineman, David; Groeneveld, Dirk; Brahman, Faeze; Timbers, Finbarr (15 December 2025). "Olmo 3". arXiv:2512.13961 [cs.CL].
  169. ↑ "Introducing Claude Opus 4.5". Anthropic. Archived from the original on 16 January 2026. Retrieved 8 January 2026.
  170. ↑ Binder, Matt (3 December 2025). "DeepSeek v3.2: What it is, how it compares to ChatGPT, how to try it". Mashable. Archived from the original on 21 April 2026. Retrieved 12 April 2026.
  171. ↑ "DeepSeek-V3.2 Release". DeepSeek API Docs. 1 December 2025. Archived from the original on 12 April 2026. Retrieved 12 April 2026.
  172. ↑ "DeepSeek-V3.2: Efficient Reasoning & Agentic AI". Hugging Face. 1 December 2025. Archived from the original on 1 April 2026. Retrieved 12 April 2026.
  173. ↑ "Advancing science and math with GPT-5.2". openai.com. Retrieved 4 January 2026.
  174. ↑ "Pushing Qwen3-Max-Thinking Beyond its Limits". Qwen. 25 January 2026. Archived from the original on 6 February 2026. Retrieved 6 February 2026. We further enhance Qwen3-Max-Thinking with two key innovations: (1) adaptive tool-use capabilities [...]; and (2) advanced test-time scaling techniques [...]. [...] We limit [parallel trajectories] and redirect saved computation to iterative self-reflection guided by a "take-experience" mechanism.
  175. ↑ "K2: An open source model that delivers frontier capabilities". Mohamed bin Zayed University of Artificial Intelligence. 5 December 2025. Archived from the original on 13 January 2026. Retrieved 10 September 2026.
  176. ↑ "K2 Think V2: a fully sovereign reasoning model". Mohamed bin Zayed University of Artificial Intelligence. 27 January 2026. Archived from the original on 8 August 2026. Retrieved 10 September 2026.
  177. ↑ "LICENSE · moonshotai/Kimi-K2.5". 1 January 2026. Archived from the original on 1 August 2026. Retrieved 1 August 2026 – via Hugging Face.
  178. ↑ Team, Kimi; Bai, Yifan; Bao, Yiping; Charles, Y.; Chen, Cheng; Chen, Guanduo; Chen, Haiting; Chen, Huarong; Chen, Jiahao (3 February 2026). "Kimi K2: Open Agentic Intelligence". arXiv:2507.20534 [cs.LG].
  179. ↑ Team, Kimi; Bai, Tongtong; Bai, Yifan; Bao, Yiping; Cai, S. H.; Cao, Yuan; Charles, Y.; Che, H. S.; Chen, Cheng (2 February 2026). "Kimi K2.5: Visual Agentic Intelligence". arXiv:2602.02276 [cs.CL].
  180. ↑ "Kimi K2.5: Chat with Kimi K2.5 for Free". Kimi K2.5. Retrieved 18 March 2026.
  181. ↑ Jiang, Ben (3 February 2026). "Compact AI model from China's StepFun outshines rivals from DeepSeek, Moonshot". South China Morning Post. Archived from the original on 4 February 2026. Retrieved 14 April 2026.
  182. ↑ "Step 3.5 Flash: Fast Enough to Think. Reliable Enough to Act". StepFun. 12 February 2026. Archived from the original on 16 April 2026. Retrieved 20 April 2026.
  183. ↑ "stepfun-ai/Step-3.5-Flash". 14 March 2026. Archived from the original on 12 April 2026. Retrieved 14 April 2026 – via Hugging Face.
  184. ↑ Mo, Liam; Goh, Brenda; Chen, Laurie (11 February 2026). D'Silva, Anil (ed.). "Chinese AI startup Zhipu releases new flagship model GLM-5". Reuters. Retrieved 17 February 2026.
  185. 1 2 "GLM-5 / README.md". 27 August 2026. Retrieved 28 August 2026 – via GitHub.
  186. 1 2 Butts, Dylan (16 February 2026). "Alibaba unveils Qwen3.5 as China's chatbot race shifts to AI agents". CNBC. Retrieved 26 February 2026.
  187. 1 2 "Qwen3.5: Towards Native Multimodal Agents". Qwen. 15 February 2026. Archived from the original on 22 February 2026. Retrieved 18 July 2026.
  188. ↑ "LICENSE · Qwen/Qwen3.5-397B-A17B". 16 February 2026. Archived from the original on 29 April 2026. Retrieved 10 April 2026 – via Hugging Face.
  189. ↑ "LICENSE · bharatgenai/Param2-17B-A2.4B-Thinking". 16 February 2026. Retrieved 12 April 2026 – via Hugging Face.
  190. ↑ "bharatgenai/Param2-17B-A2.4B-Thinking". Archived from the original on 28 February 2026. Retrieved 8 March 2026 – via Hugging Face.
  191. ↑ "sarvamai/sarvam-105b · Hugging Face". huggingface.co. Archived from the original on 8 March 2026. Retrieved 8 March 2026.
  192. ↑ Kumar, Abhijeet (19 February 2026). "Why Sarvam's new 105B model marks a shift in India's sovereign AI ambitions". Business Standard. Archived from the original on 4 March 2026. Retrieved 18 March 2026.
  193. ↑ Singh, Jagmeet (18 February 2026). "Indian AI lab Sarvam's new models are a major bet on the viability of open source AI". TechCrunch. Retrieved 18 March 2026.
  194. ↑ "Open-Sourcing Sarvam 30B and 105B". Sarvam AI. 6 March 2026. Archived from the original on 8 May 2026. Retrieved 8 May 2026.
  195. ↑ "sarvamai/sarvam-1-v0.5 · Hugging Face". huggingface.co. Retrieved 8 March 2026.
  196. ↑ Marquez, Javier (17 March 2026). "Una IA para reunir todas las funciones posibles: la apuesta de Mistral con Small 4 es hacer más con menos cosas" [An AI to bring together all possible functions: Mistral's bet with Small 4 is to do more with less]. Xataka (in Spanish). Retrieved 20 April 2026.
  197. ↑ "Introducing Mistral Small 4". Mistral AI. Archived from the original on 21 April 2026. Retrieved 20 April 2026.
  198. ↑ "Mystery AI model revealed to be Xiaomi's following suspicions it was DeepSeek's". Reuters. 18 March 2026. Retrieved 3 April 2026.
  199. ↑ "Xiaomi Launches Powerful AI Model MiMo-V2 Pro With 1 Trillion Parametres, 1 Million Token Context Window". NDTV Profit. 19 March 2026. Archived from the original on 12 April 2026. Retrieved 3 April 2026.
  200. ↑ "Qwen 3.6-Plus has been released, featuring an agent function that autonomously performs tasks". Gigazine. 3 April 2026. Retrieved 18 July 2026.
  201. ↑ "Qwen3.6-Plus: Towards Real World Agents". Qwen. 1 April 2026. Archived from the original on 14 August 2026. Retrieved 18 July 2026.
  202. ↑ Whitwam, Ryan (2 April 2026). "Google announces Gemma 4 open AI models, switches to Apache 2.0 license". Ars Technica. Archived from the original on 2 April 2026. Retrieved 3 April 2026.
  203. ↑ Mann, Tobias (2 April 2026). "Google battles Chinese open weights models with Gemma 4". Retrieved 3 April 2026.
  204. ↑ Franzen, Carl (7 April 2026). "AI joins the 8-hour work day as GLM ships 5.1 open source LLM, beating Opus 4.6 and GPT-5.4 on SWE-Bench Pro". VentureBeat. Retrieved 12 April 2026.
  205. ↑ "GLM-5.1: Towards Long-Horizon Tasks". Z.ai. Archived from the original on 10 April 2026. Retrieved 12 April 2026.
  206. ↑ "Introducing Muse Spark: Scaling Towards Personal Superintelligence". ai.meta.com. 8 April 2026. Archived from the original on 9 May 2026. Retrieved 9 May 2026.
  207. ↑ "A Chinese AI called 'Qwen3.6-35B-A3B,' which is more powerful than Gemma4, has been released as an open model". Gigazine. 17 April 2026. Retrieved 17 April 2026.
  208. ↑ "README.md · Qwen/Qwen3.6-35B-A3B". 15 April 2026. Retrieved 17 April 2026 – via Hugging Face.
  209. ↑ "LICENSE · moonshotai/Kimi-K2.6". 14 April 2026. Archived from the original on 3 August 2026. Retrieved 1 August 2026 – via Hugging Face.
  210. ↑ Gewirtz, David (21 April 2026). "Moonshot AI's new Kimi K2.6 swarms your complex tasks with 1,000 collaborating agents". ZDNET. Archived from the original on 4 May 2026. Retrieved 17 June 2026.
  211. ↑ Li, Yutong (22 April 2026). Du, Yu (ed.). 13小时编码、5天自主运行！Kimi K2.6开源“硬刚”闭源巨头，长程战力能否撕开AGI工业化新赛道？ [13 hours of coding, 5 days of autonomous operation! Kimi K2.6 open-source "goes head-on" against closed-source giants. Can its long-term capabilities carve out a new path for AGI industrialization?]. National Business Daily (in Chinese). Archived from the original on 26 August 2026. Retrieved 17 June 2026.
  212. ↑ "README.md · moonshotai/Kimi-K2.6". 23 April 2026. Retrieved 17 June 2026 – via Hugging Face.
  213. ↑ "LICENSE · Tencent-Hunyuan/Hy3-preview". Archived from the original on 9 September 2026. Retrieved 8 September 2026 – via GitHub.
  214. ↑ "Tencent Unveils Hy3 preview; Model Enhances Agent Capabilities and Real-World Usability". Tencent. 24 April 2026. Archived from the original on 8 September 2026. Retrieved 8 September 2026.
  215. ↑ "MiMo-V2.5-Pro | Xiaomi". mimo.xiaomi.com. Archived from the original on 29 April 2026. Retrieved 3 May 2026.
  216. ↑ Thomas, Prasanth Aby (28 April 2026). "Xiaomi releases MIT-licensed MiMo models for long-running AI agents". Computerworld. Archived from the original on 29 April 2026. Retrieved 6 May 2026.
  217. ↑ "XiaomiMiMo/MiMo-V2.5". Hugging Face. XiaomiMiMo. 27 April 2026. Archived from the original on 28 April 2026. Retrieved 3 May 2026.
  218. ↑ "LICENSE · mistralai/Mistral-Medium-3.5-128B". 29 April 2026. Archived from the original on 14 August 2026. Retrieved 14 August 2026 – via Hugging Face.
  219. ↑ "Mistral Medium 3.5 - Mistral AI". docs.mistral.ai. Archived from the original on 15 July 2026. Retrieved 10 July 2026.
  220. ↑ "Composer 2.5". Cursor. 18 May 2026. Archived from the original on 24 July 2026. Retrieved 24 July 2026.
  221. ↑ "Gemini 3.5: frontier intelligence with action". Google. 19 May 2026. Archived from the original on 13 July 2026. Retrieved 20 May 2026.
  222. ↑ Bhati, Divya (28 May 2026). "Alibaba AI Qwen latest model beats Google Gemini and ChatGPT in coding". India Today. Archived from the original on 21 July 2026. Retrieved 18 July 2026.
  223. ↑ "Qwen3.7: The Agent Frontier". Qwen. 20 May 2026. Archived from the original on 20 July 2026. Retrieved 18 July 2026.
  224. ↑ "Introducing Claude Opus 4.8". Anthropic. 28 May 2026. Archived from the original on 30 May 2026. Retrieved 30 May 2026.
  225. ↑ "Step 3.7 Flash". StepFun. 29 May 2026. Archived from the original on 30 May 2026. Retrieved 30 May 2026.
  226. ↑ "LICENSE · MiniMaxAI/MiniMax-M3". 12 June 2026. Retrieved 19 July 2026 – via Hugging Face.
  227. ↑ Chang, Minxiao (1 June 2026). "MiniMax debuts AI model built for long and complex coding tasks". South China Morning Post. Archived from the original on 23 August 2026. Retrieved 19 July 2026.
  228. ↑ "README.md · MiniMaxAI/MiniMax-M3". 12 June 2026. Retrieved 19 July 2026 – via Hugging Face.
  229. 1 2 "OpenMDW-1.1". OpenMDW. Archived from the original on 20 July 2026. Retrieved 19 July 2026.
  230. ↑ "NVIDIA announces its most powerful open-source model in the US, the 'Nemotron 3 Ultra,' and reports the start of mass production of its AI server, 'Vera Rubin.'". Gigazine. 1 June 2026. Retrieved 16 July 2026.
  231. ↑ "README.md · nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4". 24 June 2026. Retrieved 16 July 2026 – via Hugging Face.
  232. ↑ "Claude Fable 5 and Claude Mythos 5". Anthropic. 9 June 2026. Archived from the original on 10 June 2026. Retrieved 11 June 2026.
  233. ↑ "LICENSE · moonshotai/Kimi-K2.7-Code". 11 June 2026. Archived from the original on 1 August 2026. Retrieved 1 August 2026 – via Hugging Face.
  234. ↑ "The Chinese-made AI model 'Kimi K2.7 Code' has been released as an open model, boasting the best coding performance in the Kimi series and low token consumption". Gigazine. 15 June 2026. Retrieved 17 June 2026.
  235. ↑ "README.md · moonshotai/Kimi-K2.7-Code". 12 June 2026. Retrieved 17 June 2026 – via Hugging Face.
  236. ↑ Chang, Minxiao (15 June 2026). "Zhipu AI's stock rockets after Chinese firm launches GLM-5.2 model". South China Morning Post. Retrieved 17 June 2026.
  237. ↑ "China's Z.ai GLM-5.2 tops OpenAI's GPT 5.5 model on key benchmarks". The Economic Times. 17 June 2026. Archived from the original on 19 June 2026. Retrieved 17 June 2026.
  238. ↑ "GLM-5.2: Built for Long-Horizon Tasks". Z.ai. Archived from the original on 16 June 2026. Retrieved 16 June 2026.
  239. ↑ Franzen, Carl (29 June 2026). "Meituan open sources LongCat-2.0, the 1.6T, near-frontier agentic coding model that's been leading OpenRouter — trained entirely on Chinese chips". VentureBeat. Archived from the original on 1 July 2026. Retrieved 16 July 2026.
  240. ↑ "Introducing LongCat-2.0". LongCat. 30 June 2026. Retrieved 16 July 2026.
  241. ↑ Witteveen, Sam (6 July 2026). "Tencent's Apache-licensed Hy3 takes on GLM-5.2 at half the size — and wins everywhere except coding". VentureBeat. Retrieved 15 July 2026.
  242. ↑ "Introducing Grok 4.5". x.ai. Archived from the original on 24 July 2026. Retrieved 24 July 2026.
  243. ↑ "SpaceXAI releases Grok 4.5, which Elon describes as an 'Opus-class model'". TechCrunch. 8 July 2026. Archived from the original on 24 July 2026. Retrieved 24 July 2026.
  244. ↑ "SWE-1.7". Cognition. 8 July 2026. Retrieved 24 July 2026.
  245. ↑ Varghese, Harshita (9 July 2026). "Meta debuts Muse Spark 1.1 model with preview open to developers". Reuters. Retrieved 9 July 2026.
  246. ↑ Loizos, Connie (15 July 2026). "Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling". TechCrunch. Archived from the original on 15 July 2026. Retrieved 15 July 2026.
  247. ↑ Franzen, Carl (15 July 2026). "Thinking Machines open sources first multimodal language model, Inkling, focused on low cost and 'resistance to censorship'". VentureBeat. Retrieved 15 July 2026.
  248. ↑ Franzen, Carl (27 July 2026). "Kimi K3's full weights are here, but they're 'open' with a caveat: What enterprises should know". VentureBeat. Retrieved 1 August 2026.
  249. ↑ "Moonshot AI releases Kimi K3 model weights under custom licence: What it means for enterprises". The Indian Express. 28 July 2026. Archived from the original on 1 August 2026. Retrieved 1 August 2026.
  250. ↑ Yildirim, Ece (16 July 2026). "China Just Dropped Another Bomb on America's Frontier AI Companies". Gizmodo. Retrieved 16 July 2026.
  251. 1 2 "Qwen3.8, a Chinese AI with 2.4 trillion parameters, has been released as an open model for free, boasting performance comparable to the most advanced models from OpenAI and Anthropic". Gigazine. 13 August 2026. Retrieved 13 August 2026.
  252. 1 2 Qing Yuan (14 August 2026). 阿里开源 Qwen3.8-27B 模型，编程、办公场景表现超越 Qwen3.7-Plus [Alibaba open-sources Qwen3.8-27B model, which outperforms Qwen3.7-Plus in programming and office applications]. ITHome. Archived from the original on 15 August 2026. Retrieved 15 August 2026.
  253. ↑ "Alibaba's Qwen Unveils Preview of Flagship AI Model". Bloomberg News. 19 July 2026. Archived from the original on 19 August 2026. Retrieved 19 July 2026.
  254. 1 2 "Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber". Google. 21 July 2026. Retrieved 22 July 2026.
  255. ↑ Martin, Iain (21 July 2026). "American Open-Source Labs Think They Can Beat China's Best AI Startups". Forbes. Retrieved 23 July 2026.
  256. ↑ Nuñez, Michael (21 July 2026). "Poolside drops Laguna S 2.1, an open-weight coding model that beats rivals 10x its size". VentureBeat. Retrieved 23 July 2026.
  257. ↑ Yoon, Jeong-hoon (21 July 2026). 한국 AI의 쾌거…모티프3, 中 딥시크V4·키미K.26 어깨 나란히 [A Triumph for Korean AI... Motif 3 Stands Shoulder to Shoulder with China's DeepSeek-V4 and Kimi K2.6]. Edaily [ko] (in Korean). Archived from the original on 21 July 2026. Retrieved 14 August 2026.
  258. ↑ Kwon, Yong-man (13 August 2026). 독파모 ‘모티프 3’, 한국 1위·글로벌 10위권 평가… 오픈소스 공개 [Dokpamo's "Motif 3" Ranked No. 1 in Korea and Among the Top 10 Globally... Released as Open-Source]. IT Chosun. Retrieved 14 August 2026.
  259. ↑ "README.md · Motif-Technologies/Motif-3". 12 August 2026. Retrieved 14 August 2026 – via Hugging Face.
  260. ↑ "LICENSE · upstage/Solar-Open2-250B". 22 July 2026. Retrieved 14 August 2026 – via Hugging Face.
  261. ↑ Jhoo, Dong-chan (23 July 2026). "Upstage unveils Solar Open 2, open-weight AI model built for autonomous tasks". The Korea Times. Archived from the original on 14 August 2026. Retrieved 14 August 2026.
  262. ↑ "README.md · upstage/Solar-Open2-250B". 31 July 2026. Archived from the original on 14 August 2026. Retrieved 14 August 2026 – via Hugging Face.
  263. ↑ "Fugu". Sakana AI. Archived from the original on 24 July 2026. Retrieved 24 July 2026.
  264. ↑ Brandom, Russell (24 July 2026). "Anthropic launches Opus 5". TechCrunch. Archived from the original on 25 August 2026. Retrieved 24 July 2026.
  265. ↑ Shim, Min-gwan (29 July 2026). "SK Telecom debuts A.X K2, drives Korea's AI into industry, defense, biotech". ChosunBiz. Archived from the original on 14 August 2026. Retrieved 14 August 2026.
  266. ↑ "README.md · skt/A.X-K2". 8 August 2026. Retrieved 14 August 2026 – via Hugging Face.
  267. ↑ Franzen, Carl (30 July 2026). "Thinking Machines debuts Inkling Small open source AI model nearing performance of predecessor at about 1/4 size". VentureBeat. Retrieved 1 August 2026.
  268. ↑ Jo, He-rim (31 July 2026). "LG unveils Korea's largest AI model with 750 billion parameters". The Korea Herald. Retrieved 14 August 2026.
  269. ↑ Choi, A-ri (10 August 2026). "Four AI Teams Face Elimination in National Project Evaluation". The Chosun Daily. Retrieved 14 August 2026.
  270. ↑ "DeepSeek Doubles Down on Low-Cost AI With V4-Flash Update". MIT Sloan Management Review Middle East. 3 August 2026. Retrieved 14 August 2026.
  271. 1 2 Butts, Dylan (24 April 2026). "China's DeepSeek releases preview of long-awaited V4 model as AI race intensifies". CNBC.
  272. ↑ Baptista, Eduardo (3 August 2026). "DeepSeek's new AI model is by far the cheapest of well-known models to run, research firm says". Reuters. Retrieved 13 August 2026.
  273. ↑ Mohanty, Tamanna (10 August 2026). "Meta launches Muse Glimmer, a 30B open-weight AI model designed to run locally". Fortune India. Archived from the original on 10 August 2026. Retrieved 10 August 2026.
  274. ↑ "Gemini 3.7 Flash". Google. 13 August 2026. Archived from the original on 13 August 2026. Retrieved 13 August 2026.
  275. 1 2 "Chinese AI startup DeepSeek quietly released 'DeepSeek V4 Pro 0813'". Gigazine. 13 August 2026. Retrieved 13 August 2026.
  276. ↑ "LICENSE · tencent/Hy4-preview". 28 August 2026. Retrieved 30 August 2026 – via Hugging Face.
  277. ↑ Bhati, Divya (28 August 2026). "Tencent releases new AI model, says it beat ZAI and Moonshoot in testing". India Today. Archived from the original on 30 August 2026. Retrieved 30 August 2026.
  278. ↑ "Tencent Touts New AI Model It Claims Outperforms Z.AI, Moonshot". Bloomberg News. 28 August 2026. Retrieved 30 August 2026.
  279. 1 2 Patterson, Ben (1 September 2026). "Claude's updated Fable and Mythos models watermark their replies". PCWorld. Retrieved 2 September 2026.
  280. ↑ "Gemini 3.8 Flash". Google. 2 September 2026. Archived from the original on 2 September 2026. Retrieved 2 September 2026.
  281. ↑ Bensinger, Greg (3 September 2026). "OpenAI launches new Astra model amid growing scrutiny over agents' safety". Reuters. San Francisco. Retrieved 4 September 2026.
  282. ↑ "Introducing K2 Horizon: Frontier Performance, Radically Open". Mohamed bin Zayed University of Artificial Intelligence. 3 September 2026. Archived from the original on 9 September 2026. Retrieved 10 September 2026.
  283. ↑ Haralu, Kileuna (10 September 2026). "DeepSeek Launches V4.1-Flash to Make Large-Scale AI Inference Cheaper". Analytics India Mag. Retrieved 10 September 2026.
  284. ↑ "README.md · deepseek-ai/DeepSeek-V4.1-Flash". 10 September 2026. Retrieved 10 September 2026 – via Hugging Face.
  285. ↑ "Introducing Grok 4.7". SpaceXAI. 21 September 2026. Retrieved 30 September 2026.
  286. ↑ Wright, Webb (22 September 2026). "OpenAI Releases New GPT-6 Sol and Luna Models, Sells Them as Mini-Astras". Gizmodo. Retrieved 30 September 2026.
  287. 1 2 3 4 "MiMo-V2.6: Scaling Reinforcement Learning Toward Self-Improvement" (PDF). Hugging Face. XiaomiMiMo. 22 September 2026. Archived (PDF) from the original on 27 September 2026. Retrieved 30 September 2026.
  288. ↑ "Introducing MiMo-V2.6 series". Xiaomi MiMo. 22 September 2026. Retrieved 30 September 2026.
  289. ↑ "Anthropic unveils Claude Opus 5.5". Reuters. 22 September 2026.
  290. ↑ Capoot, Ashley (28 September 2026). "Anthropic launches cheaper AI model, its second release since CEO's call for a slowdown". CNBC. Archived from the original on 28 September 2026. Retrieved 30 September 2026.
  291. ↑ Dellinger, Aj (29 September 2026). "With No Astra to Release, OpenAI Pivots to New GPT-6.1 Sol Model". Gizmodo. Archived from the original on 1 October 2026. Retrieved 30 September 2026.
  292. ↑ Subin, Samantha; Sigalos, MacKenzie. "Google rolls out Gemini 4 Argon, its most advanced AI model". CNBC.
  293. ↑ Lellouche, Nicolas. "Mistral relance la France dans la course à l'IA et dévoile « Le Chonk », un modèle de frontière à 1000 milliards de paramètres". Numerama (in French).