LLM Context Window Growth

OpenAI
GPT-3, 2048, https://arxiv.org/pdf/2005.14165
GPT-4, 8192, https://openai.com/index/gpt-4-research/
GPT-3.5-turbo-16k, 16384, https://openai.com/index/function-calling-and-other-api-updates/
GPT-4-32k, 32768, https://openai.com/index/gpt-4-research/
GPT-4 Turbo, 128000, https://openai.com/index/new-models-and-developer-products-announced-at-devday/
GPT-4.1, 1000000, https://openai.com/index/gpt-4-1/

Google
Gemini 1.5 Pro, 1000000, https://blog.google/technology/ai/google-gemini-next-generation-model-february-2024/
Gemini 1.5 Pro, 2000000, https://ai.google.dev/gemini-api/docs/long-context

Anthropic
Claude 1, 9000, https://www.anthropic.com/news/100k-context-windows
Claude 1, 100000, https://www.anthropic.com/news/100k-context-windows
Claude 2.1, 200000, https://docs.anthropic.com/en/docs/about-claude/models/overview
Claude (1M beta), 1000000, https://www.anthropic.com/news/1m-context

Alibaba
Qwen-7B, 8192, https://github.com/QwenLM/Qwen
Qwen 1.5, 32768, https://qwen.readthedocs.io/en/v1.5/
Qwen 2.0, 32768, https://qwen.readthedocs.io/en/v2.0/
Qwen2.5, 128000, https://huggingface.co/Qwen/Qwen2.5-7B-Instruct
Qwen2.5-Turbo, 1000000, https://qwenlm.github.io/blog/qwen2.5-turbo/
Qwen2.5-1M, 1000000, https://qwenlm.github.io/blog/qwen2.5-1m/

xAI
Grok-1, 8192, https://x.ai/news/grok/model-card
Grok-1.5, 128000, https://x.ai/news/grok-1.5
Grok-3, 1000000, https://x.ai/news/grok-3

Meta
LLaMA 1, 2048, https://arxiv.org/abs/2302.13971
LLaMA 2, 4096, https://arxiv.org/pdf/2307.09288
Llama 3.1, 128000, https://ai.meta.com/blog/meta-llama-3-1/

DeepSeek
deepseek-chat / deepseek-reasoner, 128000, https://api-docs.deepseek.com/quick_start/pricing