LLM AtlasLLM AtlasSearch models

Leaderboard

structured output rankings

Updated weekly

Each 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.

Full-profile entries
245

Models with complete enough metadata and scoring coverage to be meaningfully ranked in this category.

Ranking basis
structured output

Scores combine benchmark evidence, product metadata, and cost/context signals when those fields are published.

Catalog caveat
25

Tracked models without full scoring remain in the directory and provider pages, but are not relied on for analytical ranking claims.

RankModelProviderScoreContext
#1Claude 3.7 SonnetAnthropic90200,000
#2Claude Sonnet 4.5Anthropic901,000,000
#3Claude Sonnet 4.6Anthropic901,000,000
#4GPT-5.4OpenAI901,000,000
#5Gemini 3.1 ProGoogle DeepMind891,048,576
#6GPT-5.2OpenAI891,000,000
#7Claude Opus 4.6Anthropic881,000,000
#8Claude Sonnet 4Anthropic881,000,000
#9Gemini 2.5 ProGoogle DeepMind881,048,576
#10Gemini 3.1 FlashGoogle DeepMind881,048,576
#11GPT-4oOpenAI88128,000
#12Gemini 3.0 FlashGoogle DeepMind871,048,576
#13Gemini 3.0 ProGoogle DeepMind871,048,576
#14GPT-5.3-CodexOpenAI871,000,000
#15Command R+ 2026Cohere86128,000
#16Gemini 2.5 FlashGoogle DeepMind861,048,576
#17Gemini 2.5 Pro TTSGoogle DeepMind861,048,576
#18GPT-5OpenAI861,000,000
#19GPT-5 miniOpenAI861,000,000
#20GPT-5.3 InstantOpenAI861,000,000
#21GPT-5.4 ProOpenAI861,000,000
#22Claude Haiku 4.5Anthropic85200,000
#23Gemini 2.0 FlashGoogle DeepMind851,048,576
#24GPT-4.1OpenAI851,048,576
#25GPT-5.2 ProOpenAI851,000,000
#26Mistral Large 25Mistral AI85128,000
#27Claude Opus 4.1Anthropic84200,000
#28o4-miniOpenAI84200,000
#29Claude Opus 4Anthropic83200,000
#30Gemini 2.5 Flash LiveGoogle DeepMind831,048,576
#31Gemini 2.5 Flash Native Audio PreviewGoogle DeepMind831,048,576
#32Gemini 2.5 Flash-LiteGoogle DeepMind831,048,576
#33Llama 4 MaverickMeta831,048,576
#34Gemini 3.1 Flash-LiteGoogle DeepMind821,048,576
#35GPT-4.1 miniOpenAI821,048,576
#36Sonar Reasoning ProPerplexity82200,000
#37Gemini 1.5 ProGoogle DeepMind812,097,152
#38GPT-4o-miniOpenAI81128,000
#39GPT-5 nanoOpenAI811,000,000
#40o4-mini-deep-researchOpenAI81200,000
#41warpgrep-v2Morph81128,000
#42Claude Haiku 3.5Anthropic80200,000
#43Gemini 2.0 Flash-LiteGoogle DeepMind801,048,576
#44morph-v3-fast-applyMorph80128,000
#45Nova ProAmazon Web Services80300,000
#46Sonar Deep ResearchPerplexity80200,000
#47Gemini 1.5 FlashGoogle DeepMind791,048,576
#48SonarPerplexity79128,000
#49Sonar ProPerplexity79200,000
#50Claude Sonnet 3Anthropic78200,000
#51flash-compactMorph77200,000
#52o1-miniOpenAI77128,000
#53o3OpenAI77200,000
#54Claude Haiku 3Anthropic76200,000
#55Nova LiteAmazon Web Services76300,000
#56o3-deep-researchOpenAI76200,000
#57Claude Haiku 3Anthropic75200,000
#58Claude Opus 3Anthropic75200,000
#59Gemini 1.5 Flash-8BGoogle DeepMind751,048,576
#60gpt-audioOpenAI75128,000
#61gpt-realtimeOpenAI75128,000
#62Claude Haiku 3.5Anthropic74200,000
#63Claude Haiku 4.5Anthropic74200,000
#64CodestralMistral AI74256,000
#65Codestral EmbedMistral AI7432,768
#66Doubao-Seed-1.6ByteDance / Doubao74128,000
#67Doubao-Seed-1.6-FlashByteDance / Doubao74128,000
#68Doubao-Seed-2.0-CodeByteDance / Doubao74128,000
#69Doubao-Seed-CodeByteDance / Doubao74128,000
#70gpt-oss-120bOpenAI74131,072
#71image-01MiniMax748,192
#72image-01-liveMiniMax748,192
#73Llama 3.3 70B InstructMeta74128,000
#74MiniMax-M1MiniMax74204,800
#75MiniMax-Speech-02MiniMax748,192
#76MiniMax-Text-01MiniMax74204,800
#77MiniMax-VL-01MiniMax74204,800
#78Mistral EmbedMistral AI7432,768
#79Mistral ModerationMistral AI7432,768
#80music-2.0MiniMax748,192
#81Voxtral Mini TranscribeMistral AI74131,072
#82CogVideoXZ.AI73128,000
#83CogView 4Z.AI73128,000
#84Devstral Medium 1.0Mistral AI73128,000
#85ERNIE 3.5 128KBaidu / ERNIE73128,000
#86ERNIE 4.0 Turbo 8KBaidu / ERNIE73128,000
#87ERNIE 4.5 Turbo 32KBaidu / ERNIE7332,768
#88ERNIE Functions 8KBaidu / ERNIE73128,000
#89ERNIE Speed 128KBaidu / ERNIE73128,000
#90GLM-4.5Z.AI73128,000
#91GLM-4.5VZ.AI73128,000
#92GLM-4.6Z.AI73128,000
#93GLM-4.6VZ.AI73128,000
#94GLM-4.7Z.AI73128,000
#95GLM-5Z.AI73128,000
#96GLM-ImageZ.AI73128,000
#97GLM-OCRZ.AI73128,000
#98Hunyuan CodeTencent / Hunyuan73128,000
#99Hunyuan LiteTencent / Hunyuan73128,000
#100Hunyuan StandardTencent / Hunyuan73128,000
#101Hunyuan T1Tencent / Hunyuan73256,000
#102Hunyuan T1 VisionTencent / Hunyuan73128,000
#103Hunyuan TurboSTencent / Hunyuan73128,000
#104Hunyuan TurboS LongText 128KTencent / Hunyuan73128,000
#105Kimi K2Moonshot AI / Kimi73131,072
#106Kimi K2 ThinkingMoonshot AI / Kimi73256,000
#107Kimi K2 Turbo PreviewMoonshot AI / Kimi73256,000
#108Kimi K2.5Moonshot AI / Kimi73256,000
#109Magistral Medium 1.2Mistral AI73128,000
#110MiniMax-M2MiniMax73204,800
#111MiniMax-M2.1MiniMax73204,800
#112MiniMax-M2.1-highspeedMiniMax73204,800
#113MiniMax-M2.5MiniMax73204,800
#114MiniMax-M2.5-highspeedMiniMax73204,800
#115Mistral Large 3Mistral AI73128,000
#116Mistral Medium 3.1Mistral AI73128,000
#117Mistral OCR 2505Mistral AI7332,768
#118Mistral Small 3.1Mistral AI73128,000
#119Mistral Small 3.2 OpenMistral AI73128,000
#120Nova MicroAmazon Web Services73128,000
#121o1OpenAI73200,000
#122Pixtral 12BMistral AI73131,072
#123Pixtral LargeMistral AI73131,072
#124Vidu Q1Z.AI73128,000
#125Voxtral Mini OpenMistral AI73131,072
#126Voxtral Small OpenMistral AI73131,072
#127Step-3.5-FlashStepFun72131,072
#128Claude Sonnet 3Anthropic71200,000
#129Claude Sonnet 4Anthropic711,000,000
#130Command ACohere71256,000
#131Command A ReasoningCohere71128,000
#132Command A TranslateCohere71128,000
#133Command A VisionCohere71128,000
#134Command R+Cohere71128,000
#135Command R7BCohere71128,000
#136DeepSeek-Coder-V2DeepSeek71128,000
#137DeepSeek-Math-V2DeepSeek71128,000
#138DeepSeek-OCRDeepSeek7116,384
#139DeepSeek-OCR-2DeepSeek7116,384
#140DeepSeek-R1DeepSeek71128,000
#141DeepSeek-R1-Distill-Llama-70BDeepSeek71128,000
#142DeepSeek-V2.5DeepSeek71128,000
#143DeepSeek-V3DeepSeek71128,000
#144DeepSeek-V3.1DeepSeek71128,000
#145DeepSeek-V3.1-BaseDeepSeek71128,000
#146DeepSeek-V3.2DeepSeek71128,000
#147DeepSeek-V3.2-ExpDeepSeek71128,000
#148DeepSeek-VL2-SmallDeepSeek7116,384
#149Devstral 2 OpenMistral AI71128,000
#150Devstral Small 2Mistral AI71128,000
#151Embed 4Cohere71128,000
#152Grok 3xAI71131,072
#153Grok 3 MinixAI71131,072
#154Grok 4xAI71256,000
#155Grok 4 Fast ReasoningxAI71131,072
#156grok-imagexAI71131,072
#157Jamba 3BAI21 Labs71256,000
#158Jamba LargeAI21 Labs71256,000
#159Jamba Large 1.6AI21 Labs71256,000
#160Jamba MiniAI21 Labs71256,000
#161Jamba Mini 1.6AI21 Labs71256,000
#162Jamba Mini 1.7AI21 Labs71256,000
#163Janus-Pro-7BDeepSeek7116,384
#164Magistral Small 1.2 OpenMistral AI71128,000
#165Ministral 3 14B OpenMistral AI71128,000
#166Ministral 3 3B OpenMistral AI71128,000
#167Ministral 3 8B OpenMistral AI71128,000
#168Mistral Large 3 OpenMistral AI71128,000
#169Mistral Nemo 12BMistral AI71128,000
#170phi-1Microsoft714,096
#171phi-1_5Microsoft714,096
#172phi-2Microsoft714,096
#173Phi-3-medium-4k-instructMicrosoft714,096
#174Phi-3-mini-4k-instructMicrosoft714,096
#175Phi-3-vision-128k-instructMicrosoft71128,000
#176Phi-3.5-mini-instructMicrosoft71131,072
#177Phi-3.5-MoE-instructMicrosoft71131,072
#178Phi-3.5-vision-instructMicrosoft71131,072
#179Phi-4Microsoft7116,384
#180Phi-4-mini-flash-reasoningMicrosoft71131,072
#181Phi-4-mini-instructMicrosoft71131,072
#182Phi-4-multimodal-instructMicrosoft71131,072
#183Phi-4-reasoningMicrosoft71131,072
#184Phi-4-reasoning-plusMicrosoft71131,072
#185Phi-4-reasoning-vision-15BMicrosoft71131,072
#186Phi-tiny-MoE-instructMicrosoft714,096
#187Qwen2.5-1.5B-InstructAlibaba Qwen71131,072
#188Qwen2.5-14B-InstructAlibaba Qwen71131,072
#189Qwen2.5-32B-InstructAlibaba Qwen71131,072
#190Qwen2.5-3B-InstructAlibaba Qwen71131,072
#191Qwen2.5-72B-InstructAlibaba Qwen71131,072
#192Qwen2.5-7B-InstructAlibaba Qwen71131,072
#193Qwen2.5-MaxAlibaba Qwen71131,072
#194Qwen2.5-VL-72B-InstructAlibaba Qwen71131,072
#195Qwen2.5-VL-7B-InstructAlibaba Qwen71131,072
#196Qwen3-Coder-NextAlibaba Qwen71131,072
#197Qwen3.5-0.8BAlibaba Qwen71131,072
#198Qwen3.5-122B-A10BAlibaba Qwen71131,072
#199Qwen3.5-27BAlibaba Qwen71131,072
#200Qwen3.5-2BAlibaba Qwen71131,072
#201Qwen3.5-35B-A3BAlibaba Qwen71131,072
#202Qwen3.5-397B-A17BAlibaba Qwen71131,072
#203Qwen3.5-4BAlibaba Qwen71131,072
#204Qwen3.5-9BAlibaba Qwen71131,072
#205Llama 3.1 405B InstructMeta70128,000
#206Llama 3.1 70B InstructMeta70128,000
#207Llama 4 ScoutMeta7010,485,760
#208gpt-audio-miniOpenAI69128,000
#209gpt-realtime-miniOpenAI69128,000
#210gpt-oss-20bOpenAI67131,072
#211Claude Opus 3Anthropic66200,000
#212Claude Opus 4Anthropic66200,000
#213LFM2-24B-A2BLiquid AI6632,768
#214Llama 3.1 8B InstructMeta65128,000
#215Code Llama 70B InstructMeta648,192
#216GPT Image 1OpenAI6432,768
#217Llama 3.2 90B Vision InstructMeta64128,000
#218Meta Llama 3 70B InstructMeta648,192
#219chatgpt-image-latestOpenAI6232,768
#220GPT-4o mini TranscribeOpenAI62128,000
#221GPT-4o mini TTSOpenAI62128,000
#222GPT-4o TranscribeOpenAI62128,000
#223gpt-image-1-miniOpenAI6232,768
#224MiMo-VL-7BXiaomi62131,072
#225Step3-VL-10BStepFun62131,072
#226Code Llama 34B InstructMeta618,192
#227Llama 3.2 1B InstructMeta61128,000
#228Llama 3.2 3B InstructMeta61128,000
#229Meta Llama 3 8B InstructMeta618,192
#230LFM2-8B-A1BLiquid AI6032,768
#231Llama 3.2 11B Vision InstructMeta60128,000
#232Llama Guard 4 12BMeta60131,072
#233MiMo-Audio-7BXiaomi59131,072
#234Step-Audio-R1.1StepFun58131,072
#235Llama Guard 3 11B VisionMeta56131,072
#236pplx-embed-v1-4bPerplexity558,192
#237pplx-embed-v1-0.6bPerplexity548,192
#238LFM2.5-1.2B-InstructLiquid AI53131,072
#239LFM2.5-1.2B-ThinkingLiquid AI53131,072
#240FLUX 1.1 ProBlack Forest Labs51512
#241LFM2-2.6BLiquid AI5132,768
#242FLUX 1 ProBlack Forest Labs49512
#243FLUX 1.1 Pro UltraBlack Forest Labs49512
#244Prompt Guard 86MMeta49512
#245NextStep-1.1StepFun37512

Why #1: Claude 3.7 Sonnet

A top-tier reasoning model with strong software engineering assistance and enterprise controls.

This model clears the current full-profile threshold for leaderboard methodology.

Why #2: Claude Sonnet 4.5

Anthropic's Sonnet 4.5 with 1M token context for fast frontier reasoning, coding, and long-context agent work.

This model clears the current full-profile threshold for leaderboard methodology.

Why #3: 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.