Largest Context Window

Models ranked by maximum context window size in tokens. Larger context windows allow processing more text in a single request.

Methodology: Sorted by context_window_tokens descending. All published models with context window data included.

# Model Metric
1 Gemini 1.5 Pro 2,000,000 tokens
2 Gemini 1.5 Flash 1,000,000 tokens
3 Gemini 2.0 Flash 1,000,000 tokens
4 o1 200,000 tokens
5 Claude 3 Opus 200,000 tokens
6 Claude 3 Sonnet 200,000 tokens
7 Claude 3 Haiku 200,000 tokens
8 Claude 3.5 Sonnet 200,000 tokens
9 Claude 3.5 Haiku 200,000 tokens
10 Claude Sonnet 4.6 200,000 tokens
11 Grok 2 131,072 tokens
12 GPT-4 Turbo 128,000 tokens
13 GPT-4o 128,000 tokens
14 GPT-4o mini 128,000 tokens
15 DeepSeek V3 128,000 tokens
16 Llama 3.1 405B 128,000 tokens
17 Llama 3.3 70B 128,000 tokens
18 Mistral Large 2 128,000 tokens
19 Command R+ 128,000 tokens
20 Qwen 2.5 72B 128,000 tokens
21 Claude 2 100,000 tokens
22 DeepSeek R1 64,000 tokens
23 Gemini 1.0 Pro 32,768 tokens
24 Mistral 7B 32,768 tokens
25 Mixtral 8x7B 32,768 tokens
26 GPT-3.5 Turbo 16,385 tokens
27 GPT-4 8,192 tokens
28 Llama 3 70B 8,192 tokens
29 GPT-3 (davinci-002) 4,096 tokens
30 Llama 2 70B 4,096 tokens
← All leaderboards