Context windows - verified
AI model context windows compared
70 of the 167 models in the registry accept 1M tokens or more, and the largest context window is 1,050,000 tokens (10 OpenAI models, including GPT-5.4 and GPT-5.5); 36 models charge more per token once a prompt crosses a size threshold.
Long-context pricing - Models by size - What the limits mean
- Models
- 167
- 1M+ tokens
- 70
- Largest window
- 1.05M
- Long-context pricing
- 36
Every model, largest context first
Context window and max output in tokens, as each provider states them. Output share is max output as a share of the window. Long-context pricing shows the prompt size where rates go up and the base and higher input and output price per 1M tokens.
167 models
| Model | Provider | Context windowtokens | Max outputtokens | Output shareof the window | Long-context pricingper 1M tokens, base to long | Released |
|---|---|---|---|---|---|---|
| No models match these filters. | ||||||
GPT-5.4gpt-5.4 | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $2.50 to $5 (2x)Out $15 to $22.50 (1.5x) | |
GPT-5.5gpt-5.5 | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $5 to $10 (2x)Out $30 to $45 (1.5x) | |
GPT-5.5 Progpt-5.5-pro | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $30 to $60 (2x)Out $180 to $270 (1.5x) | |
GPT-5.6 Lunagpt-5.6-luna | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $0.20 to $0.40 (2x)Out $1.20 to $1.80 (1.5x) | |
GPT-5.6 Solgpt-5.6-sol | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $4 to $8 (2x)Out $20 to $30 (1.5x) | |
GPT-5.6 Terragpt-5.6-terra | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $2 to $4 (2x)Out $12 to $18 (1.5x) | |
GPT-6 Astragpt-6-astra | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $10 to $20 (2x)Out $50 to $75 (1.5x) | |
GPT-6 Lunagpt-6-luna | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $0.10 to $0.20 (2x)Out $0.50 to $0.75 (1.5x) | |
GPT-6 Solgpt-6-sol | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $2 to $4 (2x)Out $10 to $15 (1.5x) | |
GPT-6.1 Solgpt-6.1-sol | OpenAI | 1,050,000 | 128,000 | 12% | Over 272K (272,000 tokens)In $2 to $4 (2x)Out $10 to $15 (1.5x) | |
Kimi K3kimi-k3 | Moonshot AI | 1,048,576 | 1,048,576 | 100% | -No long-context price tier | Not stated |
Gemini 2.5 Flashgemini-2.5-flash | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
Gemini 2.5 Flash-Litegemini-2.5-flash-lite | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
Gemini 2.5 Progemini-2.5-pro | 1,048,576 | 65,536 | 6.3% | Over 200K (200,000 tokens)In $1.25 to $2.50 (2x)Out $10 to $15 (1.5x) | ||
Gemini 3 Flash Previewgemini-3-flash-preview | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
Gemini 3.1 Flash-Litegemini-3.1-flash-lite | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
Gemini 3.1 Pro Previewgemini-3.1-pro-preview | 1,048,576 | 65,536 | 6.3% | Over 200K (200,000 tokens)In $2 to $4 (2x)Out $12 to $18 (1.5x) | ||
Gemini 3.5 Flash-Litegemini-3.5-flash-lite | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
Gemini 3.6 Flashgemini-3.6-flash | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
Gemini 3.8 Flashgemini-3.8-flash | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
Gemini 4 Argongemini-4-argon | 1,048,576 | 65,536 | 6.3% | -No long-context price tier | ||
GPT-4.1gpt-4.1 | OpenAI | 1,047,576 | 32,768 | 3.1% | -No long-context price tier | |
GPT-4.1 Minigpt-4.1-mini | OpenAI | 1,047,576 | 32,768 | 3.1% | -No long-context price tier | |
GPT-4.1 nanogpt-4.1-nano | OpenAI | 1,047,576 | 32,768 | 3.1% | -No long-context price tier | |
MiniMax-M3MiniMax-M3 | MiniMax | 1,000,000 | 524,288 | 52% | Over 512K (512,000 tokens)In $0.30 to $0.60 (2x)Out $1.20 to $2.40 (2x) | |
DeepSeek V4 Prodeepseek-v4-pro | DeepSeek | 1,000,000 | 393,216 | 39% | -No long-context price tier | |
DeepSeek V4.1 Flashdeepseek-flash | DeepSeek | 1,000,000 | 393,216 | 39% | -No long-context price tier | |
MiMo-V2.6-Flashmimo-v2.6-flash | Xiaomi MiMo | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
MiMo-V2.6-Promimo-v2.6-pro | Xiaomi MiMo | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
Qwen3.7-Flashqwen3.7-flash | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | Over 32K (32,000 tokens)In $0.03 to $0.10 (3.33x)Out $0.13 to $0.40 (3.08x) | |
Qwen3.7-Maxqwen3.7-max | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
Qwen3.7-Plusqwen3.7-plus | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | Over 256K (256,000 tokens)In $0.40 to $1.20 (3x)Out $1.60 to $4.80 (3x) | |
Qwen3.8-2.4T-A95Bqwen3.8-2.4t-a95b | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
Qwen3.8-27Bqwen3.8-27b | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
Qwen3.8-Flashqwen3.8-flash | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
Qwen3.8-Maxqwen3.8-max | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
Qwen3.8-Omni-Flashqwen3.8-omni-flash | Alibaba Cloud (Qwen) | 1,000,000 | 131,072 | 13% | -No long-context price tier | |
Claude Fable 5claude-fable-5 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Fable 5.1claude-fable-5-1 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Haiku 5.5claude-haiku-5-5 | Anthropic | 1,000,000 | 128,000 | 13% | Over 100K (100,000 tokens)In $0.10 to $0.50 (5x)Out $0.50 to $2.50 (5x) | |
Claude Mythos 5claude-mythos-5 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Mythos 5.1claude-mythos-5-1 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Opus 4.6claude-opus-4-6 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Opus 4.7claude-opus-4-7 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Opus 4.8claude-opus-4-8 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Opus 5claude-opus-5 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Opus 5.5claude-opus-5-5 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Sonnet 4.6claude-sonnet-4-6 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Sonnet 5claude-sonnet-5 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
Claude Sonnet 5.5claude-sonnet-5-5 | Anthropic | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
GLM-5.2glm-5.2 | Z.ai | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
GLM-5.3glm-5.3 | Z.ai | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
GLM-5.3-Flashglm-5.3-flash | Z.ai | 1,000,000 | 128,000 | 13% | -No long-context price tier | |
GLM-5.3-FlashXglm-5.3-flashx | Z.ai | 1,000,000 | 128,000 | 13% | -No long-context price tier | Not stated |
Amazon Nova 2 Liteamazon.nova-2-lite-v1:0 | Amazon | 1,000,000 | 65,536 | 6.6% | -No long-context price tier | |
Qwen3-Coder-Flashqwen3-coder-flash | Alibaba Cloud (Qwen) | 1,000,000 | 65,536 | 6.6% | Over 32K (32,000 tokens)In $0.30 to $0.50 (1.67x)Out $1.50 to $2.50 (1.67x) | |
Qwen3.5-Flashqwen3.5-flash | Alibaba Cloud (Qwen) | 1,000,000 | 65,536 | 6.6% | -No long-context price tier | |
Qwen3.5-Plusqwen3.5-plus | Alibaba Cloud (Qwen) | 1,000,000 | 65,536 | 6.6% | Over 256K (256,000 tokens)In $0.40 to $0.50 (1.25x)Out $2.40 to $3 (1.25x) | |
Qwen3.6-Flashqwen3.6-flash | Alibaba Cloud (Qwen) | 1,000,000 | 65,536 | 6.6% | Over 256K (256,000 tokens)In $0.25 to $1 (4x)Out $1.50 to $4 (2.67x) | |
Qwen3.6-Plusqwen3.6-plus | Alibaba Cloud (Qwen) | 1,000,000 | 65,536 | 6.6% | Over 256K (256,000 tokens)In $0.50 to $2 (4x)Out $3 to $6 (2x) | |
Hy4 previewhy4-preview | Tencent | 1,000,000 | 64,000 | 6.4% | -No long-context price tier | |
Step 5 Previewstep-5-preview | StepFun | 1,000,000 | 64,000 | 6.4% | -No long-context price tier | Not stated |
Qwen-Flashqwen-flash | Alibaba Cloud (Qwen) | 1,000,000 | 32,768 | 3.3% | Over 256K (256,000 tokens)In $0.05 to $0.25 (5x)Out $0.40 to $2 (5x) | |
Qwen-Plusqwen-plus | Alibaba Cloud (Qwen) | 1,000,000 | 32,768 | 3.3% | Over 256K (256,000 tokens)In $0.40 to $1.20 (3x)Out $1.20 to $3.60 (3x) | Not stated |
Amazon Nova Premieramazon.nova-premier-v1:0 | Amazon | 1,000,000 | 10,000 | 1% | -No long-context price tier | |
Grok 4.20grok-4.20-0309-reasoning | xAI | 1,000,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $1.25 to $2.50 (2x)Out $2.50 to $5 (2x) | |
Grok 4.20 (Non-Reasoning)grok-4.20-0309-non-reasoning | xAI | 1,000,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $1.25 to $2.50 (2x)Out $2.50 to $5 (2x) | |
Grok 4.20 Multi-Agent Betagrok-4.20-multi-agent-0309 | xAI | 1,000,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $1.25 to $2.50 (2x)Out $2.50 to $5 (2x) | |
Grok 4.3grok-4.3 | xAI | 1,000,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $1.25 to $2.50 (2x)Out $2.50 to $5 (2x) | Not stated |
Mistral Large 4mistral-large-4 | Mistral AI | 1,000,000 | Not stated | -Not computed | -No long-context price tier | |
Grok 4.5grok-4.5 | xAI | 500,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $2 to $4 (2x)Out $6 to $12 (2x) | |
Grok 4.6grok-4.6 | xAI | 500,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $2 to $4 (2x)Out $6 to $12 (2x) | |
Grok 4.7grok-4.7 | xAI | 500,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $2 to $4 (2x)Out $6 to $12 (2x) | |
GPT-5 Progpt-5-pro | OpenAI | 400,000 | 272,000 | 68% | -No long-context price tier | |
Chat Latestchat-latest | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | Not stated |
GPT-5gpt-5 | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5 Minigpt-5-mini | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5 nanogpt-5-nano | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5.1gpt-5.1 | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5.2gpt-5.2 | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5.3-Codexgpt-5.3-codex | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5.4 Minigpt-5.4-mini | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5.4 nanogpt-5.4-nano | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
GPT-5.6 Cybergpt-5.6-cyber | OpenAI | 400,000 | 128,000 | 32% | -No long-context price tier | |
Amazon Nova Liteamazon.nova-lite-v1:0 | Amazon | 300,000 | 10,000 | 3.3% | -No long-context price tier | |
Amazon Nova Proamazon.nova-pro-v1:0 | Amazon | 300,000 | 10,000 | 3.3% | -No long-context price tier | |
Qwen3.5-122B-A10Bqwen3.5-122b-a10b | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3.5-27Bqwen3.5-27b | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3.5-35B-A3Bqwen3.5-35b-a3b | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3.5-397B-A17Bqwen3.5-397b-a17b | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3.5-Omni-Flashqwen3.5-omni-flash | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3.5-Omni-Plusqwen3.5-omni-plus | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3.6-27Bqwen3.6-27b | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3.6-35B-A3Bqwen3.6-35b-a3b | Alibaba Cloud (Qwen) | 262,144 | 65,536 | 25% | -No long-context price tier | |
Qwen3-VL-Plusqwen3-vl-plus | Alibaba Cloud (Qwen) | 262,144 | 32,768 | 13% | Over 32K (32,000 tokens)In $0.20 to $0.30 (1.5x)Out $1.60 to $2.40 (1.5x) | |
Kimi K2.6kimi-k2.6 | Moonshot AI | 262,144 | Not stated | -Not computed | -No long-context price tier | Not stated |
Kimi K2.7 Codekimi-k2.7-code | Moonshot AI | 262,144 | Not stated | -Not computed | -No long-context price tier | Not stated |
Kimi K2.7 Code HighSpeedkimi-k2.7-code-highspeed | Moonshot AI | 262,144 | Not stated | -Not computed | -No long-context price tier | Not stated |
Mercury 2.5mercury-2.5 | Inception | 260,000 | 65,536 | 25% | -No long-context price tier | |
Dola Seed 2.1 Turbodola-seed-2-1-turbo-260628 | ByteDance Seed | 256,000 | 256,000 | 100% | -No long-context price tier | Not stated |
Hy3hy3 | Tencent | 256,000 | 128,000 | 50% | -No long-context price tier | |
Seed 2.0 Liteseed-2-0-lite-260428 | ByteDance Seed | 256,000 | 128,000 | 50% | Over 128K (128,000 tokens)In $0.25 to $0.50 (2x)Out $2 to $4 (2x) | Not stated |
Seed 2.0 Miniseed-2-0-mini-260428 | ByteDance Seed | 256,000 | 128,000 | 50% | Over 128K (128,000 tokens)In $0.10 to $0.20 (2x)Out $0.40 to $0.80 (2x) | Not stated |
Seed 2.0 Proseed-2-0-pro-260328 | ByteDance Seed | 256,000 | 128,000 | 50% | Over 128K (128,000 tokens)In $0.50 to $1 (2x)Out $3 to $6 (2x) | Not stated |
Command Acommand-a-03-2025 | Cohere | 256,000 | 8,000 | 3.1% | -No long-context price tier | |
Jamba Large 1.7jamba-large-1.7 | AI21 Labs | 256,000 | 4,096 | 1.6% | -No long-context price tier | |
Jamba2 Minijamba-mini-2 | AI21 Labs | 256,000 | 4,096 | 1.6% | -No long-context price tier | |
Grok Build 0.1grok-build-0.1 | xAI | 256,000 | Not stated | -Not computed | Over 200K (200,000 tokens)In $1 to $2 (2x)Out $2 to $4 (2x) | |
Ministral 3 14Bministral-14b-2512 | Mistral AI | 256,000 | Not stated | -Not computed | -No long-context price tier | |
Ministral 3 3Bministral-3b-2512 | Mistral AI | 256,000 | Not stated | -Not computed | -No long-context price tier | |
Ministral 3 8Bministral-8b-2512 | Mistral AI | 256,000 | Not stated | -Not computed | -No long-context price tier | |
Mistral Large 3mistral-large-2512 | Mistral AI | 256,000 | Not stated | -Not computed | -No long-context price tier | |
Mistral Medium 3.5mistral-medium-3-5 | Mistral AI | 256,000 | Not stated | -Not computed | -No long-context price tier | |
Mistral Small 4mistral-small-2603 | Mistral AI | 256,000 | Not stated | -Not computed | -No long-context price tier | |
Step 3.5 Flashstep-3.5-flash | StepFun | 256,000 | Not stated | -Not computed | -No long-context price tier | Not stated |
Step 3.5 Flash 2603step-3.5-flash-2603 | StepFun | 256,000 | Not stated | -Not computed | -No long-context price tier | Not stated |
Step 3.7 Flashstep-3.7-flash | StepFun | 256,000 | Not stated | -Not computed | -No long-context price tier | |
MiniMax-M2.5MiniMax-M2.5 | MiniMax | 204,800 | 204,800 | 100% | -No long-context price tier | |
MiniMax-M2.5-highspeedMiniMax-M2.5-highspeed | MiniMax | 204,800 | 204,800 | 100% | -No long-context price tier | |
MiniMax-M2.7MiniMax-M2.7 | MiniMax | 204,800 | 204,800 | 100% | -No long-context price tier | |
MiniMax-M2.7-highspeedMiniMax-M2.7-highspeed | MiniMax | 204,800 | 204,800 | 100% | -No long-context price tier | |
GLM-4.6glm-4.6 | Z.ai | 200,000 | 128,000 | 64% | -No long-context price tier | |
GLM-4.7glm-4.7 | Z.ai | 200,000 | 128,000 | 64% | -No long-context price tier | |
GLM-4.7-FlashXglm-4.7-flashx | Z.ai | 200,000 | 128,000 | 64% | -No long-context price tier | Not stated |
GLM-5glm-5 | Z.ai | 200,000 | 128,000 | 64% | -No long-context price tier | |
GLM-5.1glm-5.1 | Z.ai | 200,000 | 128,000 | 64% | -No long-context price tier | |
o1o1 | OpenAI | 200,000 | 100,000 | 50% | -No long-context price tier | |
o1-proo1-pro | OpenAI | 200,000 | 100,000 | 50% | -No long-context price tier | |
o3o3 | OpenAI | 200,000 | 100,000 | 50% | -No long-context price tier | |
o3-minio3-mini | OpenAI | 200,000 | 100,000 | 50% | -No long-context price tier | |
o3-proo3-pro | OpenAI | 200,000 | 100,000 | 50% | -No long-context price tier | |
o4-minio4-mini | OpenAI | 200,000 | 100,000 | 50% | -No long-context price tier | |
Claude Haiku 4.5claude-haiku-4-5-20251001 | Anthropic | 200,000 | 64,000 | 32% | -No long-context price tier | |
Claude Opus 4.5claude-opus-4-5-20251101 | Anthropic | 200,000 | 64,000 | 32% | -No long-context price tier | |
Claude Sonnet 4.5claude-sonnet-4-5-20250929 | Anthropic | 200,000 | 64,000 | 32% | -No long-context price tier | |
Gemini Robotics ER 2 Previewgemini-robotics-er-2-preview | 131,072 | 65,536 | 50% | -No long-context price tier | ||
GLM-4.5glm-4.5 | Z.ai | 128,000 | 96,000 | 75% | -No long-context price tier | |
GLM-4.5-Airglm-4.5-air | Z.ai | 128,000 | 96,000 | 75% | -No long-context price tier | |
ERNIE 5.0ernie-5.0 | Baidu | 128,000 | 65,536 | 51% | Over 32K (32,000 tokens)In $0.84 to $1.40 (1.67x)Out $3.38 to $5.60 (1.66x) | Not stated |
ERNIE 5.1ernie-5.1 | Baidu | 128,000 | 65,536 | 51% | Over 32K (32,000 tokens)In $0.56 to $0.84 (1.5x)Out $2.53 to $3.10 (1.23x) | |
Mercury 2mercury-2 | Inception | 128,000 | 50,000 | 39% | -No long-context price tier | |
ERNIE 4.5 Turbo VLernie-4.5-turbo-vl | Baidu | 128,000 | 16,384 | 13% | -No long-context price tier | Not stated |
GPT-4ogpt-4o | OpenAI | 128,000 | 16,384 | 13% | -No long-context price tier | |
GPT-4o Minigpt-4o-mini | OpenAI | 128,000 | 16,384 | 13% | -No long-context price tier | |
GLM-4-32B-0414-128Kglm-4-32b-0414-128k | Z.ai | 128,000 | 16,000 | 13% | -No long-context price tier | Not stated |
ERNIE 4.5 Turbo (128K)ernie-4.5-turbo-128k | Baidu | 128,000 | 12,288 | 9.6% | -No long-context price tier | |
Amazon Nova Microamazon.nova-micro-v1:0 | Amazon | 128,000 | 10,000 | 7.8% | -No long-context price tier | |
GPT-4 Turbogpt-4-turbo | OpenAI | 128,000 | 4,096 | 3.2% | -No long-context price tier | |
Aya Expanse 32Bc4ai-aya-expanse-32b | Cohere | 128,000 | 4,000 | 3.1% | -No long-context price tier | |
Command R 08-2024command-r-08-2024 | Cohere | 128,000 | 4,000 | 3.1% | -No long-context price tier | |
Command R+ 08-2024command-r-plus-08-2024 | Cohere | 128,000 | 4,000 | 3.1% | -No long-context price tier | |
Command R7Bcommand-r7b-12-2024 | Cohere | 128,000 | 4,000 | 3.1% | -No long-context price tier | |
Codestral 25.08codestral-2508 | Mistral AI | 128,000 | Not stated | -Not computed | -No long-context price tier | |
GLM-4.6Vglm-4.6v | Z.ai | 128,000 | Not stated | -Not computed | -No long-context price tier | |
Qwen-Maxqwen-max | Alibaba Cloud (Qwen) | 32,768 | 8,192 | 25% | -No long-context price tier | |
ERNIE 4.5 Turbo (32K)ernie-4.5-turbo-32k | Baidu | 32,000 | 12,288 | 38% | -No long-context price tier | |
ERNIE 4.5 Turbo VL (32K)ernie-4.5-turbo-vl-32k | Baidu | 32,000 | 12,288 | 38% | -No long-context price tier | Not stated |
Mercury Edit 2mercury-edit-2 | Inception | 32,000 | 8,192 | 26% | -No long-context price tier | |
Step-1o Turbo Visionstep-1o-turbo-vision | StepFun | 32,000 | Not stated | -Not computed | -No long-context price tier | Not stated |
GPT-3.5 Turbogpt-3.5-turbo | OpenAI | 16,385 | 4,096 | 25% | -No long-context price tier | Not stated |
Qwen-MT-Flashqwen-mt-flash | Alibaba Cloud (Qwen) | 16,384 | 8,192 | 50% | -No long-context price tier | |
Qwen-MT-Liteqwen-mt-lite | Alibaba Cloud (Qwen) | 16,384 | 8,192 | 50% | -No long-context price tier | |
GPT-4gpt-4 | OpenAI | 8,192 | 8,192 | 100% | -No long-context price tier | Not stated |
Hy-MT2-Litehy-mt2-lite | Tencent | 8,000 | 4,000 | 50% | -No long-context price tier | Not stated |
Hy-MT2-Plushy-mt2-plus | Tencent | 8,000 | 4,000 | 50% | -No long-context price tier | Not stated |
Hy-MT2-Prohy-mt2-pro | Tencent | 8,000 | 4,000 | 50% | -No long-context price tier | Not stated |
GPT-3.5 Turbo Instructgpt-3.5-turbo-instruct | OpenAI | 4,096 | 4,096 | 100% | -No long-context price tier | Not stated |
Where a big prompt gets more expensive
Long-context pricing
36 of the 167 models, from 8 providers, bill the whole request at a higher rate once the prompt passes a token threshold. Thresholds run from 32,000 to 512,000 tokens, most often 200,000 (10 models). Input rates rise 1.25x to 5x (most often 2x) and output rates 1.23x to 5x (most often 1.5x).
| Model | Thresholdprompt over, tokens | Base-rate shareof the window | Inputper 1M, base to long | Outputper 1M, base to long | Cache readper 1M, base to long |
|---|---|---|---|---|---|
| Qwen3.7-FlashAlibaba Cloud (Qwen) - 1M window | 32,000 | 3.2% | $0.03 to $0.10 (3.33x) | $0.13 to $0.40 (3.08x) | Not stated above threshold |
| ERNIE 5.0Baidu - 128K window | 32,000 | 25% | $0.84 to $1.40 (1.67x) | $3.38 to $5.60 (1.66x) | No cache price |
| Qwen3-Coder-FlashAlibaba Cloud (Qwen) - 1M window | 32,000 | 3.2% | $0.30 to $0.50 (1.67x) | $1.50 to $2.50 (1.67x) | Not stated above threshold |
| ERNIE 5.1Baidu - 128K window | 32,000 | 25% | $0.56 to $0.84 (1.5x) | $2.53 to $3.10 (1.23x) | No cache price |
| Qwen3-VL-PlusAlibaba Cloud (Qwen) - 262K window | 32,000 | 12% | $0.20 to $0.30 (1.5x) | $1.60 to $2.40 (1.5x) | Not stated above threshold |
| Claude Haiku 5.5Anthropic - 1M window | 100,000 | 10% | $0.10 to $0.50 (5x) | $0.50 to $2.50 (5x) | $0.01 to $0.05 (5x) |
| Seed 2.0 LiteByteDance Seed - 256K window | 128,000 | 50% | $0.25 to $0.50 (2x) | $2 to $4 (2x) | $0.05 to $0.10 (2x) |
| Seed 2.0 MiniByteDance Seed - 256K window | 128,000 | 50% | $0.10 to $0.20 (2x) | $0.40 to $0.80 (2x) | $0.02 to $0.04 (2x) |
| Seed 2.0 ProByteDance Seed - 256K window | 128,000 | 50% | $0.50 to $1 (2x) | $3 to $6 (2x) | $0.10 to $0.20 (2x) |
| Gemini 2.5 ProGoogle - 1.05M window | 200,000 | 19% | $1.25 to $2.50 (2x) | $10 to $15 (1.5x) | $0.125 to $0.25 (2x) |
| Gemini 3.1 Pro PreviewGoogle - 1.05M window | 200,000 | 19% | $2 to $4 (2x) | $12 to $18 (1.5x) | $0.20 to $0.40 (2x) |
| Grok 4.20xAI - 1M window | 200,000 | 20% | $1.25 to $2.50 (2x) | $2.50 to $5 (2x) | $0.20 to $0.40 (2x) |
| Grok 4.20 (Non-Reasoning)xAI - 1M window | 200,000 | 20% | $1.25 to $2.50 (2x) | $2.50 to $5 (2x) | $0.20 to $0.40 (2x) |
| Grok 4.20 Multi-Agent BetaxAI - 1M window | 200,000 | 20% | $1.25 to $2.50 (2x) | $2.50 to $5 (2x) | $0.20 to $0.40 (2x) |
| Grok 4.3xAI - 1M window | 200,000 | 20% | $1.25 to $2.50 (2x) | $2.50 to $5 (2x) | $0.20 to $0.40 (2x) |
| Grok 4.5xAI - 500K window | 200,000 | 40% | $2 to $4 (2x) | $6 to $12 (2x) | $0.30 to $0.60 (2x) |
| Grok 4.6xAI - 500K window | 200,000 | 40% | $2 to $4 (2x) | $6 to $12 (2x) | $0.50 to $1 (2x) |
| Grok 4.7xAI - 500K window | 200,000 | 40% | $2 to $4 (2x) | $6 to $12 (2x) | $0.50 to $1 (2x) |
| Grok Build 0.1xAI - 256K window | 200,000 | 78% | $1 to $2 (2x) | $2 to $4 (2x) | $0.20 to $0.40 (2x) |
| Qwen-FlashAlibaba Cloud (Qwen) - 1M window | 256,000 | 26% | $0.05 to $0.25 (5x) | $0.40 to $2 (5x) | Not stated above threshold |
| Qwen3.6-FlashAlibaba Cloud (Qwen) - 1M window | 256,000 | 26% | $0.25 to $1 (4x) | $1.50 to $4 (2.67x) | Not stated above threshold |
| Qwen3.6-PlusAlibaba Cloud (Qwen) - 1M window | 256,000 | 26% | $0.50 to $2 (4x) | $3 to $6 (2x) | Not stated above threshold |
| Qwen-PlusAlibaba Cloud (Qwen) - 1M window | 256,000 | 26% | $0.40 to $1.20 (3x) | $1.20 to $3.60 (3x) | Not stated above threshold |
| Qwen3.7-PlusAlibaba Cloud (Qwen) - 1M window | 256,000 | 26% | $0.40 to $1.20 (3x) | $1.60 to $4.80 (3x) | Not stated above threshold |
| Qwen3.5-PlusAlibaba Cloud (Qwen) - 1M window | 256,000 | 26% | $0.40 to $0.50 (1.25x) | $2.40 to $3 (1.25x) | Not stated above threshold |
| GPT-5.4OpenAI - 1.05M window | 272,000 | 26% | $2.50 to $5 (2x) | $15 to $22.50 (1.5x) | $0.25 to $0.50 (2x) |
| GPT-5.5OpenAI - 1.05M window | 272,000 | 26% | $5 to $10 (2x) | $30 to $45 (1.5x) | $0.50 to $1 (2x) |
| GPT-5.5 ProOpenAI - 1.05M window | 272,000 | 26% | $30 to $60 (2x) | $180 to $270 (1.5x) | No cache price |
| GPT-5.6 LunaOpenAI - 1.05M window | 272,000 | 26% | $0.20 to $0.40 (2x) | $1.20 to $1.80 (1.5x) | $0.02 to $0.04 (2x) |
| GPT-5.6 SolOpenAI - 1.05M window | 272,000 | 26% | $4 to $8 (2x) | $20 to $30 (1.5x) | $0.40 to $0.80 (2x) |
| GPT-5.6 TerraOpenAI - 1.05M window | 272,000 | 26% | $2 to $4 (2x) | $12 to $18 (1.5x) | $0.20 to $0.40 (2x) |
| GPT-6 AstraOpenAI - 1.05M window | 272,000 | 26% | $10 to $20 (2x) | $50 to $75 (1.5x) | $1 to $2 (2x) |
| GPT-6 LunaOpenAI - 1.05M window | 272,000 | 26% | $0.10 to $0.20 (2x) | $0.50 to $0.75 (1.5x) | $0.01 to $0.02 (2x) |
| GPT-6 SolOpenAI - 1.05M window | 272,000 | 26% | $2 to $4 (2x) | $10 to $15 (1.5x) | $0.20 to $0.40 (2x) |
| GPT-6.1 SolOpenAI - 1.05M window | 272,000 | 26% | $2 to $4 (2x) | $10 to $15 (1.5x) | $0.10 to $0.20 (2x) |
| MiniMax-M3MiniMax - 1M window | 512,000 | 51% | $0.30 to $0.60 (2x) | $1.20 to $2.40 (2x) | $0.06 to $0.12 (2x) |
The multiplier is the long-context rate divided by the base rate, rounded to two decimals. Base-rate share is the threshold as a share of the context window: the rest of the window is billed at the higher rate. Some providers bill more than two tiers; the dataset records the first step up, and the model page note lists the rest.
Models by context window size
1M tokens or more70 models
- GPT-5.41,050,000
- GPT-5.51,050,000
- GPT-5.5 Pro1,050,000
- GPT-5.6 Luna1,050,000
- GPT-5.6 Sol1,050,000
- GPT-5.6 Terra1,050,000
- GPT-6 Astra1,050,000
- GPT-6 Luna1,050,000
- GPT-6 Sol1,050,000
- GPT-6.1 Sol1,050,000
- Kimi K31,048,576
- Gemini 2.5 Flash1,048,576
- Gemini 2.5 Flash-Lite1,048,576
- Gemini 2.5 Pro1,048,576
- Gemini 3 Flash Preview1,048,576
- Gemini 3.1 Flash-Lite1,048,576
- Gemini 3.1 Pro Preview1,048,576
- Gemini 3.5 Flash-Lite1,048,576
- Gemini 3.6 Flash1,048,576
- Gemini 3.8 Flash1,048,576
- Gemini 4 Argon1,048,576
- GPT-4.11,047,576
- GPT-4.1 Mini1,047,576
- GPT-4.1 nano1,047,576
- MiniMax-M31,000,000
- DeepSeek V4 Pro1,000,000
- DeepSeek V4.1 Flash1,000,000
- MiMo-V2.6-Flash1,000,000
- MiMo-V2.6-Pro1,000,000
- Qwen3.7-Flash1,000,000
- Qwen3.7-Max1,000,000
- Qwen3.7-Plus1,000,000
- Qwen3.8-2.4T-A95B1,000,000
- Qwen3.8-27B1,000,000
- Qwen3.8-Flash1,000,000
- Qwen3.8-Max1,000,000
- Qwen3.8-Omni-Flash1,000,000
- Claude Fable 51,000,000
- Claude Fable 5.11,000,000
- Claude Haiku 5.51,000,000
- Claude Mythos 51,000,000
- Claude Mythos 5.11,000,000
- Claude Opus 4.61,000,000
- Claude Opus 4.71,000,000
- Claude Opus 4.81,000,000
- Claude Opus 51,000,000
- Claude Opus 5.51,000,000
- Claude Sonnet 4.61,000,000
- Claude Sonnet 51,000,000
- Claude Sonnet 5.51,000,000
- GLM-5.21,000,000
- GLM-5.31,000,000
- GLM-5.3-Flash1,000,000
- GLM-5.3-FlashX1,000,000
- Amazon Nova 2 Lite1,000,000
- Qwen3-Coder-Flash1,000,000
- Qwen3.5-Flash1,000,000
- Qwen3.5-Plus1,000,000
- Qwen3.6-Flash1,000,000
- Qwen3.6-Plus1,000,000
- Hy4 preview1,000,000
- Step 5 Preview1,000,000
- Qwen-Flash1,000,000
- Qwen-Plus1,000,000
- Amazon Nova Premier1,000,000
- Grok 4.201,000,000
- Grok 4.20 (Non-Reasoning)1,000,000
- Grok 4.20 Multi-Agent Beta1,000,000
- Grok 4.31,000,000
- Mistral Large 41,000,000
400K to 1M tokens14 models
- Grok 4.5500,000
- Grok 4.6500,000
- Grok 4.7500,000
- GPT-5 Pro400,000
- Chat Latest400,000
- GPT-5400,000
- GPT-5 Mini400,000
- GPT-5 nano400,000
- GPT-5.1400,000
- GPT-5.2400,000
- GPT-5.3-Codex400,000
- GPT-5.4 Mini400,000
- GPT-5.4 nano400,000
- GPT-5.6 Cyber400,000
128K to 400K tokens70 models
- Amazon Nova Lite300,000
- Amazon Nova Pro300,000
- Qwen3.5-122B-A10B262,144
- Qwen3.5-27B262,144
- Qwen3.5-35B-A3B262,144
- Qwen3.5-397B-A17B262,144
- Qwen3.5-Omni-Flash262,144
- Qwen3.5-Omni-Plus262,144
- Qwen3.6-27B262,144
- Qwen3.6-35B-A3B262,144
- Qwen3-VL-Plus262,144
- Kimi K2.6262,144
- Kimi K2.7 Code262,144
- Kimi K2.7 Code HighSpeed262,144
- Mercury 2.5260,000
- Dola Seed 2.1 Turbo256,000
- Hy3256,000
- Seed 2.0 Lite256,000
- Seed 2.0 Mini256,000
- Seed 2.0 Pro256,000
- Command A256,000
- Jamba Large 1.7256,000
- Jamba2 Mini256,000
- Grok Build 0.1256,000
- Ministral 3 14B256,000
- Ministral 3 3B256,000
- Ministral 3 8B256,000
- Mistral Large 3256,000
- Mistral Medium 3.5256,000
- Mistral Small 4256,000
- Step 3.5 Flash256,000
- Step 3.5 Flash 2603256,000
- Step 3.7 Flash256,000
- MiniMax-M2.5204,800
- MiniMax-M2.5-highspeed204,800
- MiniMax-M2.7204,800
- MiniMax-M2.7-highspeed204,800
- GLM-4.6200,000
- GLM-4.7200,000
- GLM-4.7-FlashX200,000
- GLM-5200,000
- GLM-5.1200,000
- o1200,000
- o1-pro200,000
- o3200,000
- o3-mini200,000
- o3-pro200,000
- o4-mini200,000
- Claude Haiku 4.5200,000
- Claude Opus 4.5200,000
- Claude Sonnet 4.5200,000
- Gemini Robotics ER 2 Preview131,072
- GLM-4.5128,000
- GLM-4.5-Air128,000
- ERNIE 5.0128,000
- ERNIE 5.1128,000
- Mercury 2128,000
- ERNIE 4.5 Turbo VL128,000
- GPT-4o128,000
- GPT-4o Mini128,000
- GLM-4-32B-0414-128K128,000
- ERNIE 4.5 Turbo (128K)128,000
- Amazon Nova Micro128,000
- GPT-4 Turbo128,000
- Aya Expanse 32B128,000
- Command R 08-2024128,000
- Command R+ 08-2024128,000
- Command R7B128,000
- Codestral 25.08128,000
- GLM-4.6V128,000
Under 128K tokens13 models
- Qwen-Max32,768
- ERNIE 4.5 Turbo (32K)32,000
- ERNIE 4.5 Turbo VL (32K)32,000
- Mercury Edit 232,000
- Step-1o Turbo Vision32,000
- GPT-3.5 Turbo16,385
- Qwen-MT-Flash16,384
- Qwen-MT-Lite16,384
- GPT-48,192
- Hy-MT2-Lite8,000
- Hy-MT2-Plus8,000
- Hy-MT2-Pro8,000
- GPT-3.5 Turbo Instruct4,096
Explainer
What a context window is
The numbers above are each provider's own limits. Here is how to read them before you size a prompt.
- The context window
- The most tokens one request can hold: system prompt, conversation history, documents, tool definitions and tool results, plus the response the model writes. Anything past it is rejected or has to be trimmed. See the context window glossary entry.
- Input vs output limits
- Max output caps the response alone, and it is usually far smaller than the window: the median here is 100,000 tokens, 13% of the window. The output also counts against the window, so a prompt that nearly fills it leaves little room for the answer. Where a provider documents a max output equal to the whole window (Kimi K3 and several MiniMax models here), that is a ceiling on one shared budget, not room for a full prompt plus a full-length answer. Reasoning models can spend part of the output budget on thinking tokens before the visible reply.
- A bigger window is not the same as using it well
- Accepting 1M tokens does not mean recalling all of them equally. Models tend to use the start and end of a long prompt better than the middle (lost in the middle), and quality can drop well before the advertised limit (effective context length). Long prompts also cost more and take longer, and some models charge a higher rate past a threshold. Retrieval, caching and a trimmed prompt usually beat filling the window.
FAQ
Which AI model has the largest context window?
The largest context window in the registry is 1,050,000 tokens, shared by 10 OpenAI models, including GPT-5.4 and GPT-5.5. The median across all 167 models is 400,000 tokens, and the smallest is 4,096.
How many AI models have a 1M token context window?
70 of the 167 models accept 1,000,000 tokens or more, from 14 providers. The most come from Alibaba Cloud (Qwen) (15), Anthropic (13), OpenAI (13) and Google (10).
What is the difference between context window and max output?
The context window is the total number of tokens a request can hold, prompt and response together; max output caps the response alone. Of the 143 models whose provider states a max output, the median is 100,000 tokens, 13% of the window. The largest is 1,048,576 tokens (Kimi K3).
Do long prompts cost more?
For 36 of the 167 models, yes: once the prompt passes a threshold the whole request is billed at higher rates. Thresholds run from 32,000 to 512,000 tokens, most often 200,000 (10 models). Input rates rise 1.25x to 5x, most often 2x.
Verified against each provider's own pricing and model pages. Limits and prices change when a provider updates a model or repoints an alias; open a model for its sources before you rely on a number.
Newsletter
Get model price changes as they ship
One email per week. New model releases, price cuts and what they mean for your bill.
Keep exploring
Model API pricing and specs
First-party prices per 1M tokens and sources for every model.
Model API capabilities
Tool calling, structured outputs, vision, caching and reasoning per model.
Model knowledge cutoffs
Provider-stated cutoff dates, release dates and the gap between them.
AI cost calculator
Estimate what a workload costs across models.