Leaderboard
long context rankings
Updated weeklyEach leaderboard uses transparent weighted scoring, current model context, and supporting analysis to help teams interpret the results with confidence. Only full-profile entries appear in rankings; broader catalog records remain available elsewhere on the site when only source-backed metadata is currently available.
Models with complete enough metadata and scoring coverage to be meaningfully ranked in this category.
Scores combine benchmark evidence, product metadata, and cost/context signals when those fields are published.
Tracked models without full scoring remain in the directory and provider pages, but are not relied on for analytical ranking claims.
| Rank | Model | Provider | Score | Context |
|---|---|---|---|---|
| #1 | GPT-5.4 | OpenAI | 96 | 1,000,000 |
| #2 | Claude Sonnet 4.6 | Anthropic | 95 | 1,000,000 |
| #3 | Gemini 3.1 Pro | Google DeepMind | 95 | 1,048,576 |
| #4 | GPT-5.2 | OpenAI | 95 | 1,000,000 |
| #5 | Claude 3.7 Sonnet | Anthropic | 94 | 200,000 |
| #6 | Claude Sonnet 4.5 | Anthropic | 94 | 1,000,000 |
| #7 | Gemini 2.5 Pro | Google DeepMind | 94 | 1,048,576 |
| #8 | Gemini 3.0 Pro | Google DeepMind | 94 | 1,048,576 |
| #9 | Gemini 3.1 Flash | Google DeepMind | 94 | 1,048,576 |
| #10 | GPT-4o | OpenAI | 94 | 128,000 |
| #11 | GPT-5.3-Codex | OpenAI | 94 | 1,000,000 |
| #12 | Claude Opus 4.6 | Anthropic | 93 | 1,000,000 |
| #13 | Claude Sonnet 4 | Anthropic | 93 | 1,000,000 |
| #14 | Gemini 2.5 Flash | Google DeepMind | 93 | 1,048,576 |
| #15 | Gemini 2.5 Pro TTS | Google DeepMind | 93 | 1,048,576 |
| #16 | Gemini 3.0 Flash | Google DeepMind | 93 | 1,048,576 |
| #17 | GPT-5 | OpenAI | 93 | 1,000,000 |
| #18 | GPT-5 mini | OpenAI | 93 | 1,000,000 |
| #19 | GPT-5.2 Pro | OpenAI | 93 | 1,000,000 |
| #20 | GPT-5.3 Instant | OpenAI | 93 | 1,000,000 |
| #21 | GPT-5.4 Pro | OpenAI | 93 | 1,000,000 |
| #22 | Claude Haiku 4.5 | Anthropic | 92 | 200,000 |
| #23 | Command R+ 2026 | Cohere | 92 | 128,000 |
| #24 | Gemini 2.0 Flash | Google DeepMind | 92 | 1,048,576 |
| #25 | GPT-4.1 | OpenAI | 92 | 1,048,576 |
| #26 | Llama 4 Maverick | Meta | 92 | 1,048,576 |
| #27 | Mistral Large 25 | Mistral AI | 92 | 128,000 |
| #28 | Claude Opus 4.1 | Anthropic | 91 | 200,000 |
| #29 | Gemini 2.5 Flash Live | Google DeepMind | 91 | 1,048,576 |
| #30 | Gemini 2.5 Flash Native Audio Preview | Google DeepMind | 91 | 1,048,576 |
| #31 | Gemini 2.5 Flash-Lite | Google DeepMind | 91 | 1,048,576 |
| #32 | Gemini 3.1 Flash-Lite | Google DeepMind | 91 | 1,048,576 |
| #33 | o4-mini | OpenAI | 91 | 200,000 |
| #34 | Claude Opus 4 | Anthropic | 90 | 200,000 |
| #35 | Gemini 1.5 Pro | Google DeepMind | 90 | 2,097,152 |
| #36 | GPT-4.1 mini | OpenAI | 90 | 1,048,576 |
| #37 | GPT-5 nano | OpenAI | 90 | 1,000,000 |
| #38 | o4-mini-deep-research | OpenAI | 90 | 200,000 |
| #39 | Sonar Reasoning Pro | Perplexity | 90 | 200,000 |
| #40 | Claude Haiku 3.5 | Anthropic | 89 | 200,000 |
| #41 | Gemini 1.5 Flash | Google DeepMind | 89 | 1,048,576 |
| #42 | Gemini 2.0 Flash-Lite | Google DeepMind | 89 | 1,048,576 |
| #43 | GPT-4o-mini | OpenAI | 89 | 128,000 |
| #44 | morph-v3-fast-apply | Morph | 89 | 128,000 |
| #45 | Nova Pro | Amazon Web Services | 89 | 300,000 |
| #46 | Sonar Deep Research | Perplexity | 89 | 200,000 |
| #47 | Sonar Pro | Perplexity | 89 | 200,000 |
| #48 | warpgrep-v2 | Morph | 89 | 128,000 |
| #49 | Llama 3.3 70B Instruct | Meta | 88 | 128,000 |
| #50 | o1-mini | OpenAI | 88 | 128,000 |
| #51 | o3 | OpenAI | 88 | 200,000 |
| #52 | Sonar | Perplexity | 88 | 128,000 |
| #53 | Claude Sonnet 3 | Anthropic | 87 | 200,000 |
| #54 | flash-compact | Morph | 87 | 200,000 |
| #55 | gpt-oss-120b | OpenAI | 87 | 131,072 |
| #56 | Nova Lite | Amazon Web Services | 87 | 300,000 |
| #57 | o3-deep-research | OpenAI | 87 | 200,000 |
| #58 | Claude Haiku 3 | Anthropic | 86 | 200,000 |
| #59 | Claude Haiku 3 | Anthropic | 86 | 200,000 |
| #60 | Claude Haiku 3.5 | Anthropic | 86 | 200,000 |
| #61 | Claude Haiku 4.5 | Anthropic | 86 | 200,000 |
| #62 | Claude Opus 3 | Anthropic | 86 | 200,000 |
| #63 | Codestral | Mistral AI | 86 | 256,000 |
| #64 | Doubao-Seed-1.6 | ByteDance / Doubao | 86 | 128,000 |
| #65 | Doubao-Seed-1.6-Flash | ByteDance / Doubao | 86 | 128,000 |
| #66 | Doubao-Seed-2.0-Code | ByteDance / Doubao | 86 | 128,000 |
| #67 | Doubao-Seed-Code | ByteDance / Doubao | 86 | 128,000 |
| #68 | Gemini 1.5 Flash-8B | Google DeepMind | 86 | 1,048,576 |
| #69 | gpt-audio | OpenAI | 86 | 128,000 |
| #70 | gpt-realtime | OpenAI | 86 | 128,000 |
| #71 | Llama 4 Scout | Meta | 86 | 10,485,760 |
| #72 | MiniMax-M1 | MiniMax | 86 | 204,800 |
| #73 | MiniMax-Text-01 | MiniMax | 86 | 204,800 |
| #74 | MiniMax-VL-01 | MiniMax | 86 | 204,800 |
| #75 | Step-3.5-Flash | StepFun | 86 | 131,072 |
| #76 | Voxtral Mini Transcribe | Mistral AI | 86 | 131,072 |
| #77 | CogVideoX | Z.AI | 85 | 128,000 |
| #78 | CogView 4 | Z.AI | 85 | 128,000 |
| #79 | DeepSeek-Coder-V2 | DeepSeek | 85 | 128,000 |
| #80 | DeepSeek-Math-V2 | DeepSeek | 85 | 128,000 |
| #81 | DeepSeek-R1 | DeepSeek | 85 | 128,000 |
| #82 | DeepSeek-R1-Distill-Llama-70B | DeepSeek | 85 | 128,000 |
| #83 | DeepSeek-V2.5 | DeepSeek | 85 | 128,000 |
| #84 | DeepSeek-V3 | DeepSeek | 85 | 128,000 |
| #85 | DeepSeek-V3.1 | DeepSeek | 85 | 128,000 |
| #86 | DeepSeek-V3.1-Base | DeepSeek | 85 | 128,000 |
| #87 | DeepSeek-V3.2 | DeepSeek | 85 | 128,000 |
| #88 | DeepSeek-V3.2-Exp | DeepSeek | 85 | 128,000 |
| #89 | Devstral 2 Open | Mistral AI | 85 | 128,000 |
| #90 | Devstral Medium 1.0 | Mistral AI | 85 | 128,000 |
| #91 | Devstral Small 2 | Mistral AI | 85 | 128,000 |
| #92 | ERNIE 3.5 128K | Baidu / ERNIE | 85 | 128,000 |
| #93 | ERNIE 4.0 Turbo 8K | Baidu / ERNIE | 85 | 128,000 |
| #94 | ERNIE Functions 8K | Baidu / ERNIE | 85 | 128,000 |
| #95 | ERNIE Speed 128K | Baidu / ERNIE | 85 | 128,000 |
| #96 | GLM-4.5 | Z.AI | 85 | 128,000 |
| #97 | GLM-4.5V | Z.AI | 85 | 128,000 |
| #98 | GLM-4.6 | Z.AI | 85 | 128,000 |
| #99 | GLM-4.6V | Z.AI | 85 | 128,000 |
| #100 | GLM-4.7 | Z.AI | 85 | 128,000 |
| #101 | GLM-5 | Z.AI | 85 | 128,000 |
| #102 | GLM-Image | Z.AI | 85 | 128,000 |
| #103 | GLM-OCR | Z.AI | 85 | 128,000 |
| #104 | Hunyuan Code | Tencent / Hunyuan | 85 | 128,000 |
| #105 | Hunyuan Lite | Tencent / Hunyuan | 85 | 128,000 |
| #106 | Hunyuan Standard | Tencent / Hunyuan | 85 | 128,000 |
| #107 | Hunyuan T1 | Tencent / Hunyuan | 85 | 256,000 |
| #108 | Hunyuan T1 Vision | Tencent / Hunyuan | 85 | 128,000 |
| #109 | Hunyuan TurboS | Tencent / Hunyuan | 85 | 128,000 |
| #110 | Hunyuan TurboS LongText 128K | Tencent / Hunyuan | 85 | 128,000 |
| #111 | Jamba 3B | AI21 Labs | 85 | 256,000 |
| #112 | Jamba Large | AI21 Labs | 85 | 256,000 |
| #113 | Jamba Large 1.6 | AI21 Labs | 85 | 256,000 |
| #114 | Jamba Mini | AI21 Labs | 85 | 256,000 |
| #115 | Jamba Mini 1.6 | AI21 Labs | 85 | 256,000 |
| #116 | Jamba Mini 1.7 | AI21 Labs | 85 | 256,000 |
| #117 | Kimi K2 | Moonshot AI / Kimi | 85 | 131,072 |
| #118 | Kimi K2 Thinking | Moonshot AI / Kimi | 85 | 256,000 |
| #119 | Kimi K2 Turbo Preview | Moonshot AI / Kimi | 85 | 256,000 |
| #120 | Kimi K2.5 | Moonshot AI / Kimi | 85 | 256,000 |
| #121 | Llama 3.1 405B Instruct | Meta | 85 | 128,000 |
| #122 | Llama 3.1 70B Instruct | Meta | 85 | 128,000 |
| #123 | Magistral Medium 1.2 | Mistral AI | 85 | 128,000 |
| #124 | Magistral Small 1.2 Open | Mistral AI | 85 | 128,000 |
| #125 | MiniMax-M2 | MiniMax | 85 | 204,800 |
| #126 | MiniMax-M2.1 | MiniMax | 85 | 204,800 |
| #127 | MiniMax-M2.1-highspeed | MiniMax | 85 | 204,800 |
| #128 | MiniMax-M2.5 | MiniMax | 85 | 204,800 |
| #129 | MiniMax-M2.5-highspeed | MiniMax | 85 | 204,800 |
| #130 | Ministral 3 14B Open | Mistral AI | 85 | 128,000 |
| #131 | Ministral 3 3B Open | Mistral AI | 85 | 128,000 |
| #132 | Ministral 3 8B Open | Mistral AI | 85 | 128,000 |
| #133 | Mistral Large 3 | Mistral AI | 85 | 128,000 |
| #134 | Mistral Large 3 Open | Mistral AI | 85 | 128,000 |
| #135 | Mistral Medium 3.1 | Mistral AI | 85 | 128,000 |
| #136 | Mistral Nemo 12B | Mistral AI | 85 | 128,000 |
| #137 | Mistral Small 3.1 | Mistral AI | 85 | 128,000 |
| #138 | Mistral Small 3.2 Open | Mistral AI | 85 | 128,000 |
| #139 | Nova Micro | Amazon Web Services | 85 | 128,000 |
| #140 | o1 | OpenAI | 85 | 200,000 |
| #141 | Phi-3-vision-128k-instruct | Microsoft | 85 | 128,000 |
| #142 | Phi-3.5-mini-instruct | Microsoft | 85 | 131,072 |
| #143 | Phi-3.5-MoE-instruct | Microsoft | 85 | 131,072 |
| #144 | Phi-3.5-vision-instruct | Microsoft | 85 | 131,072 |
| #145 | Phi-4-mini-flash-reasoning | Microsoft | 85 | 131,072 |
| #146 | Phi-4-mini-instruct | Microsoft | 85 | 131,072 |
| #147 | Phi-4-multimodal-instruct | Microsoft | 85 | 131,072 |
| #148 | Phi-4-reasoning | Microsoft | 85 | 131,072 |
| #149 | Phi-4-reasoning-plus | Microsoft | 85 | 131,072 |
| #150 | Phi-4-reasoning-vision-15B | Microsoft | 85 | 131,072 |
| #151 | Pixtral 12B | Mistral AI | 85 | 131,072 |
| #152 | Pixtral Large | Mistral AI | 85 | 131,072 |
| #153 | Qwen2.5-1.5B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #154 | Qwen2.5-14B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #155 | Qwen2.5-32B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #156 | Qwen2.5-3B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #157 | Qwen2.5-72B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #158 | Qwen2.5-7B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #159 | Qwen2.5-Max | Alibaba Qwen | 85 | 131,072 |
| #160 | Qwen2.5-VL-72B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #161 | Qwen2.5-VL-7B-Instruct | Alibaba Qwen | 85 | 131,072 |
| #162 | Qwen3-Coder-Next | Alibaba Qwen | 85 | 131,072 |
| #163 | Qwen3.5-0.8B | Alibaba Qwen | 85 | 131,072 |
| #164 | Qwen3.5-122B-A10B | Alibaba Qwen | 85 | 131,072 |
| #165 | Qwen3.5-27B | Alibaba Qwen | 85 | 131,072 |
| #166 | Qwen3.5-2B | Alibaba Qwen | 85 | 131,072 |
| #167 | Qwen3.5-35B-A3B | Alibaba Qwen | 85 | 131,072 |
| #168 | Qwen3.5-397B-A17B | Alibaba Qwen | 85 | 131,072 |
| #169 | Qwen3.5-4B | Alibaba Qwen | 85 | 131,072 |
| #170 | Qwen3.5-9B | Alibaba Qwen | 85 | 131,072 |
| #171 | Vidu Q1 | Z.AI | 85 | 128,000 |
| #172 | Voxtral Mini Open | Mistral AI | 85 | 131,072 |
| #173 | Voxtral Small Open | Mistral AI | 85 | 131,072 |
| #174 | Claude Sonnet 3 | Anthropic | 84 | 200,000 |
| #175 | Claude Sonnet 4 | Anthropic | 84 | 1,000,000 |
| #176 | Command A | Cohere | 84 | 256,000 |
| #177 | Command A Reasoning | Cohere | 84 | 128,000 |
| #178 | Command A Translate | Cohere | 84 | 128,000 |
| #179 | Command A Vision | Cohere | 84 | 128,000 |
| #180 | Command R+ | Cohere | 84 | 128,000 |
| #181 | Command R7B | Cohere | 84 | 128,000 |
| #182 | Embed 4 | Cohere | 84 | 128,000 |
| #183 | Grok 3 | xAI | 84 | 131,072 |
| #184 | Grok 3 Mini | xAI | 84 | 131,072 |
| #185 | Grok 4 | xAI | 84 | 256,000 |
| #186 | Grok 4 Fast Reasoning | xAI | 84 | 131,072 |
| #187 | grok-image | xAI | 84 | 131,072 |
| #188 | gpt-audio-mini | OpenAI | 83 | 128,000 |
| #189 | gpt-oss-20b | OpenAI | 83 | 131,072 |
| #190 | gpt-realtime-mini | OpenAI | 83 | 128,000 |
| #191 | Llama 3.1 8B Instruct | Meta | 82 | 128,000 |
| #192 | Llama 3.2 90B Vision Instruct | Meta | 82 | 128,000 |
| #193 | Claude Opus 3 | Anthropic | 81 | 200,000 |
| #194 | Claude Opus 4 | Anthropic | 81 | 200,000 |
| #195 | Codestral Embed | Mistral AI | 81 | 32,768 |
| #196 | ERNIE 4.5 Turbo 32K | Baidu / ERNIE | 81 | 32,768 |
| #197 | Mistral Embed | Mistral AI | 81 | 32,768 |
| #198 | Mistral Moderation | Mistral AI | 81 | 32,768 |
| #199 | Mistral OCR 2505 | Mistral AI | 81 | 32,768 |
| #200 | Llama 3.2 1B Instruct | Meta | 80 | 128,000 |
| #201 | Llama 3.2 3B Instruct | Meta | 80 | 128,000 |
| #202 | MiMo-VL-7B | Xiaomi | 80 | 131,072 |
| #203 | Step3-VL-10B | StepFun | 80 | 131,072 |
| #204 | Llama 3.2 11B Vision Instruct | Meta | 79 | 128,000 |
| #205 | MiMo-Audio-7B | Xiaomi | 79 | 131,072 |
| #206 | GPT-4o mini Transcribe | OpenAI | 78 | 128,000 |
| #207 | GPT-4o mini TTS | OpenAI | 78 | 128,000 |
| #208 | GPT-4o Transcribe | OpenAI | 78 | 128,000 |
| #209 | LFM2-24B-A2B | Liquid AI | 78 | 32,768 |
| #210 | Step-Audio-R1.1 | StepFun | 78 | 131,072 |
| #211 | DeepSeek-OCR | DeepSeek | 77 | 16,384 |
| #212 | DeepSeek-OCR-2 | DeepSeek | 77 | 16,384 |
| #213 | DeepSeek-VL2-Small | DeepSeek | 77 | 16,384 |
| #214 | Janus-Pro-7B | DeepSeek | 77 | 16,384 |
| #215 | Phi-4 | Microsoft | 77 | 16,384 |
| #216 | image-01 | MiniMax | 76 | 8,192 |
| #217 | image-01-live | MiniMax | 76 | 8,192 |
| #218 | Llama Guard 4 12B | Meta | 76 | 131,072 |
| #219 | MiniMax-Speech-02 | MiniMax | 76 | 8,192 |
| #220 | music-2.0 | MiniMax | 76 | 8,192 |
| #221 | LFM2-8B-A1B | Liquid AI | 75 | 32,768 |
| #222 | LFM2.5-1.2B-Instruct | Liquid AI | 75 | 131,072 |
| #223 | LFM2.5-1.2B-Thinking | Liquid AI | 75 | 131,072 |
| #224 | GPT Image 1 | OpenAI | 74 | 32,768 |
| #225 | Llama Guard 3 11B Vision | Meta | 74 | 131,072 |
| #226 | chatgpt-image-latest | OpenAI | 73 | 32,768 |
| #227 | gpt-image-1-mini | OpenAI | 73 | 32,768 |
| #228 | Code Llama 70B Instruct | Meta | 71 | 8,192 |
| #229 | Meta Llama 3 70B Instruct | Meta | 71 | 8,192 |
| #230 | phi-1 | Microsoft | 71 | 4,096 |
| #231 | phi-1_5 | Microsoft | 71 | 4,096 |
| #232 | phi-2 | Microsoft | 71 | 4,096 |
| #233 | Phi-3-medium-4k-instruct | Microsoft | 71 | 4,096 |
| #234 | Phi-3-mini-4k-instruct | Microsoft | 71 | 4,096 |
| #235 | Phi-tiny-MoE-instruct | Microsoft | 71 | 4,096 |
| #236 | Code Llama 34B Instruct | Meta | 69 | 8,192 |
| #237 | LFM2-2.6B | Liquid AI | 69 | 32,768 |
| #238 | Meta Llama 3 8B Instruct | Meta | 69 | 8,192 |
| #239 | pplx-embed-v1-4b | Perplexity | 63 | 8,192 |
| #240 | pplx-embed-v1-0.6b | Perplexity | 62 | 8,192 |
| #241 | FLUX 1.1 Pro | Black Forest Labs | 50 | 512 |
| #242 | FLUX 1 Pro | Black Forest Labs | 49 | 512 |
| #243 | FLUX 1.1 Pro Ultra | Black Forest Labs | 49 | 512 |
| #244 | Prompt Guard 86M | Meta | 47 | 512 |
| #245 | NextStep-1.1 | StepFun | 43 | 512 |
Why #1: GPT-5.4
OpenAI's GPT-5.4, the most capable and efficient frontier model for professional work. First general-purpose model with native computer-use capabilities. Combines industry-leading coding from GPT-5.3-Codex with improved agentic workflows.
This model clears the current full-profile threshold for leaderboard methodology.
Why #2: Claude Sonnet 4.6
Anthropic's current Sonnet tier for fast frontier reasoning, coding, and long-context agent work.
This model clears the current full-profile threshold for leaderboard methodology.
Why #3: Gemini 3.1 Pro
Google's Gemini 3.1 Pro, designed for complex tasks where simple answers aren't enough. Released Feb 2026 with enhanced reasoning and multimodal capabilities.
This model clears the current full-profile threshold for leaderboard methodology.