LLM AtlasLLM AtlasSearch models

Leaderboard

open source 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
open source

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
#1Llama 4 MaverickMeta911,048,576
#2gpt-oss-120bOpenAI85131,072
#3Llama 3.3 70B InstructMeta85128,000
#4Step-3.5-FlashStepFun84131,072
#5Llama 3.1 70B InstructMeta82128,000
#6Llama 4 ScoutMeta8210,485,760
#7Pixtral 12BMistral AI82131,072
#8Voxtral Mini OpenMistral AI82131,072
#9Voxtral Small OpenMistral AI82131,072
#10DeepSeek-Coder-V2DeepSeek81128,000
#11DeepSeek-Math-V2DeepSeek81128,000
#12DeepSeek-R1DeepSeek81128,000
#13DeepSeek-R1-Distill-Llama-70BDeepSeek81128,000
#14DeepSeek-V2.5DeepSeek81128,000
#15DeepSeek-V3DeepSeek81128,000
#16DeepSeek-V3.1DeepSeek81128,000
#17DeepSeek-V3.1-BaseDeepSeek81128,000
#18DeepSeek-V3.2DeepSeek81128,000
#19DeepSeek-V3.2-ExpDeepSeek81128,000
#20Devstral 2 OpenMistral AI81128,000
#21Devstral Small 2Mistral AI81128,000
#22Jamba 3BAI21 Labs81256,000
#23Jamba LargeAI21 Labs81256,000
#24Jamba Large 1.6AI21 Labs81256,000
#25Jamba MiniAI21 Labs81256,000
#26Jamba Mini 1.6AI21 Labs81256,000
#27Jamba Mini 1.7AI21 Labs81256,000
#28Llama 3.1 405B InstructMeta81128,000
#29Magistral Small 1.2 OpenMistral AI81128,000
#30Ministral 3 14B OpenMistral AI81128,000
#31Ministral 3 3B OpenMistral AI81128,000
#32Ministral 3 8B OpenMistral AI81128,000
#33Mistral Large 3 OpenMistral AI81128,000
#34Mistral Nemo 12BMistral AI81128,000
#35Phi-3-vision-128k-instructMicrosoft81128,000
#36Phi-3.5-mini-instructMicrosoft81131,072
#37Phi-3.5-MoE-instructMicrosoft81131,072
#38Phi-3.5-vision-instructMicrosoft81131,072
#39Phi-4-mini-flash-reasoningMicrosoft81131,072
#40Phi-4-mini-instructMicrosoft81131,072
#41Phi-4-multimodal-instructMicrosoft81131,072
#42Phi-4-reasoningMicrosoft81131,072
#43Phi-4-reasoning-plusMicrosoft81131,072
#44Phi-4-reasoning-vision-15BMicrosoft81131,072
#45Qwen2.5-1.5B-InstructAlibaba Qwen81131,072
#46Qwen2.5-14B-InstructAlibaba Qwen81131,072
#47Qwen2.5-32B-InstructAlibaba Qwen81131,072
#48Qwen2.5-3B-InstructAlibaba Qwen81131,072
#49Qwen2.5-72B-InstructAlibaba Qwen81131,072
#50Qwen2.5-7B-InstructAlibaba Qwen81131,072
#51Qwen2.5-MaxAlibaba Qwen81131,072
#52Qwen2.5-VL-72B-InstructAlibaba Qwen81131,072
#53Qwen2.5-VL-7B-InstructAlibaba Qwen81131,072
#54Qwen3-Coder-NextAlibaba Qwen81131,072
#55Qwen3.5-0.8BAlibaba Qwen81131,072
#56Qwen3.5-122B-A10BAlibaba Qwen81131,072
#57Qwen3.5-27BAlibaba Qwen81131,072
#58Qwen3.5-2BAlibaba Qwen81131,072
#59Qwen3.5-35B-A3BAlibaba Qwen81131,072
#60Qwen3.5-397B-A17BAlibaba Qwen81131,072
#61Qwen3.5-4BAlibaba Qwen81131,072
#62Qwen3.5-9BAlibaba Qwen81131,072
#63DeepSeek-OCRDeepSeek7916,384
#64DeepSeek-OCR-2DeepSeek7916,384
#65DeepSeek-VL2-SmallDeepSeek7916,384
#66gpt-oss-20bOpenAI79131,072
#67Janus-Pro-7BDeepSeek7916,384
#68Phi-4Microsoft7916,384
#69Llama 3.1 8B InstructMeta78128,000
#70phi-1Microsoft784,096
#71phi-1_5Microsoft784,096
#72phi-2Microsoft784,096
#73Phi-3-medium-4k-instructMicrosoft784,096
#74Phi-3-mini-4k-instructMicrosoft784,096
#75Phi-tiny-MoE-instructMicrosoft784,096
#76LFM2-24B-A2BLiquid AI7632,768
#77Llama 3.2 90B Vision InstructMeta76128,000
#78Code Llama 70B InstructMeta758,192
#79Llama 3.2 1B InstructMeta75128,000
#80Llama 3.2 3B InstructMeta75128,000
#81Meta Llama 3 70B InstructMeta758,192
#82Step3-VL-10BStepFun75131,072
#83Llama 3.2 11B Vision InstructMeta74128,000
#84MiMo-VL-7BXiaomi74131,072
#85Code Llama 34B InstructMeta738,192
#86Gemini 3.1 FlashGoogle DeepMind731,048,576
#87Meta Llama 3 8B InstructMeta738,192
#88Gemini 2.5 FlashGoogle DeepMind721,048,576
#89Gemini 3.0 FlashGoogle DeepMind721,048,576
#90Gemini 3.1 ProGoogle DeepMind721,048,576
#91GPT-5.2OpenAI721,000,000
#92GPT-5.4OpenAI721,000,000
#93MiMo-Audio-7BXiaomi72131,072
#94Claude 3.7 SonnetAnthropic71200,000
#95Claude Sonnet 4.5Anthropic711,000,000
#96Claude Sonnet 4.6Anthropic711,000,000
#97Gemini 2.0 FlashGoogle DeepMind711,048,576
#98Gemini 2.5 ProGoogle DeepMind711,048,576
#99Gemini 3.0 ProGoogle DeepMind711,048,576
#100GPT-5 miniOpenAI711,000,000
#101GPT-5.3 InstantOpenAI711,000,000
#102GPT-5.3-CodexOpenAI711,000,000
#103LFM2-8B-A1BLiquid AI7132,768
#104Step-Audio-R1.1StepFun71131,072
#105Claude Haiku 4.5Anthropic70200,000
#106Claude Sonnet 4Anthropic701,000,000
#107Gemini 2.5 Pro TTSGoogle DeepMind701,048,576
#108GPT-4.1OpenAI701,048,576
#109GPT-4oOpenAI70128,000
#110GPT-5OpenAI701,000,000
#111Mistral Large 25Mistral AI70128,000
#112o4-miniOpenAI70200,000
#113Command R+ 2026Cohere69128,000
#114Gemini 2.5 Flash LiveGoogle DeepMind691,048,576
#115Gemini 2.5 Flash Native Audio PreviewGoogle DeepMind691,048,576
#116Gemini 2.5 Flash-LiteGoogle DeepMind691,048,576
#117Gemini 3.1 Flash-LiteGoogle DeepMind691,048,576
#118GPT-4.1 miniOpenAI691,048,576
#119morph-v3-fast-applyMorph69128,000
#120warpgrep-v2Morph69128,000
#121Claude Opus 4.6Anthropic681,000,000
#122GPT-5 nanoOpenAI681,000,000
#123flash-compactMorph67200,000
#124Gemini 1.5 FlashGoogle DeepMind671,048,576
#125Gemini 1.5 ProGoogle DeepMind672,097,152
#126Gemini 2.0 Flash-LiteGoogle DeepMind671,048,576
#127GPT-4o-miniOpenAI67128,000
#128LFM2.5-1.2B-InstructLiquid AI67131,072
#129LFM2.5-1.2B-ThinkingLiquid AI67131,072
#130Claude Haiku 3.5Anthropic66200,000
#131GPT-5.2 ProOpenAI661,000,000
#132GPT-5.4 ProOpenAI661,000,000
#133Llama Guard 4 12BMeta66131,072
#134Nova ProAmazon Web Services66300,000
#135o4-mini-deep-researchOpenAI66200,000
#136Sonar Reasoning ProPerplexity66200,000
#137LFM2-2.6BLiquid AI6532,768
#138SonarPerplexity65128,000
#139Claude Haiku 3Anthropic64200,000
#140Nova LiteAmazon Web Services64300,000
#141o1-miniOpenAI64128,000
#142Sonar Deep ResearchPerplexity64200,000
#143Claude Haiku 3Anthropic63200,000
#144Claude Haiku 3.5Anthropic63200,000
#145Claude Haiku 4.5Anthropic63200,000
#146Claude Sonnet 3Anthropic63200,000
#147CodestralMistral AI63256,000
#148Codestral EmbedMistral AI6332,768
#149Doubao-Seed-1.6ByteDance / Doubao63128,000
#150Doubao-Seed-1.6-FlashByteDance / Doubao63128,000
#151Doubao-Seed-2.0-CodeByteDance / Doubao63128,000
#152Doubao-Seed-CodeByteDance / Doubao63128,000
#153Gemini 1.5 Flash-8BGoogle DeepMind631,048,576
#154Llama Guard 3 11B VisionMeta63131,072
#155MiniMax-M1MiniMax63204,800
#156MiniMax-Text-01MiniMax63204,800
#157MiniMax-VL-01MiniMax63204,800
#158Mistral EmbedMistral AI6332,768
#159Mistral ModerationMistral AI6332,768
#160Sonar ProPerplexity63200,000
#161Voxtral Mini TranscribeMistral AI63131,072
#162Claude Opus 4Anthropic62200,000
#163Claude Opus 4.1Anthropic62200,000
#164CogVideoXZ.AI62128,000
#165CogView 4Z.AI62128,000
#166Devstral Medium 1.0Mistral AI62128,000
#167ERNIE 3.5 128KBaidu / ERNIE62128,000
#168ERNIE 4.0 Turbo 8KBaidu / ERNIE62128,000
#169ERNIE 4.5 Turbo 32KBaidu / ERNIE6232,768
#170ERNIE Functions 8KBaidu / ERNIE62128,000
#171ERNIE Speed 128KBaidu / ERNIE62128,000
#172GLM-4.5Z.AI62128,000
#173GLM-4.5VZ.AI62128,000
#174GLM-4.6Z.AI62128,000
#175GLM-4.6VZ.AI62128,000
#176GLM-4.7Z.AI62128,000
#177GLM-5Z.AI62128,000
#178GLM-ImageZ.AI62128,000
#179GLM-OCRZ.AI62128,000
#180Hunyuan CodeTencent / Hunyuan62128,000
#181Hunyuan LiteTencent / Hunyuan62128,000
#182Hunyuan StandardTencent / Hunyuan62128,000
#183Hunyuan T1Tencent / Hunyuan62256,000
#184Hunyuan T1 VisionTencent / Hunyuan62128,000
#185Hunyuan TurboSTencent / Hunyuan62128,000
#186Hunyuan TurboS LongText 128KTencent / Hunyuan62128,000
#187Kimi K2Moonshot AI / Kimi62131,072
#188Kimi K2 ThinkingMoonshot AI / Kimi62256,000
#189Kimi K2 Turbo PreviewMoonshot AI / Kimi62256,000
#190Kimi K2.5Moonshot AI / Kimi62256,000
#191Magistral Medium 1.2Mistral AI62128,000
#192Mistral Large 3Mistral AI62128,000
#193Mistral Medium 3.1Mistral AI62128,000
#194Mistral OCR 2505Mistral AI6232,768
#195Mistral Small 3.1Mistral AI62128,000
#196Mistral Small 3.2 OpenMistral AI62128,000
#197Pixtral LargeMistral AI62131,072
#198Vidu Q1Z.AI62128,000
#199image-01MiniMax618,192
#200image-01-liveMiniMax618,192
#201MiniMax-M2MiniMax61204,800
#202MiniMax-M2.1MiniMax61204,800
#203MiniMax-M2.1-highspeedMiniMax61204,800
#204MiniMax-M2.5MiniMax61204,800
#205MiniMax-M2.5-highspeedMiniMax61204,800
#206MiniMax-Speech-02MiniMax618,192
#207music-2.0MiniMax618,192
#208Nova MicroAmazon Web Services61128,000
#209Claude Sonnet 3Anthropic59200,000
#210Claude Sonnet 4Anthropic591,000,000
#211Command ACohere59256,000
#212Command A ReasoningCohere59128,000
#213Command A TranslateCohere59128,000
#214Command A VisionCohere59128,000
#215Command R+Cohere59128,000
#216Command R7BCohere59128,000
#217Embed 4Cohere59128,000
#218gpt-audioOpenAI59128,000
#219gpt-realtimeOpenAI59128,000
#220Grok 3xAI59131,072
#221Grok 3 MinixAI59131,072
#222Grok 4xAI59256,000
#223Grok 4 Fast ReasoningxAI59131,072
#224grok-imagexAI59131,072
#225o3OpenAI59200,000
#226o3-deep-researchOpenAI59200,000
#227Claude Opus 3Anthropic56200,000
#228o1OpenAI56200,000
#229gpt-audio-miniOpenAI55128,000
#230gpt-realtime-miniOpenAI55128,000
#231Prompt Guard 86MMeta53512
#232Claude Opus 3Anthropic52200,000
#233Claude Opus 4Anthropic52200,000
#234GPT Image 1OpenAI4932,768
#235GPT-4o mini TranscribeOpenAI49128,000
#236GPT-4o mini TTSOpenAI49128,000
#237GPT-4o TranscribeOpenAI49128,000
#238chatgpt-image-latestOpenAI4832,768
#239gpt-image-1-miniOpenAI4832,768
#240NextStep-1.1StepFun48512
#241pplx-embed-v1-0.6bPerplexity438,192
#242pplx-embed-v1-4bPerplexity438,192
#243FLUX 1.1 ProBlack Forest Labs39512
#244FLUX 1 ProBlack Forest Labs37512
#245FLUX 1.1 Pro UltraBlack Forest Labs37512

Why #1: Llama 4 Maverick

Meta's 17Bx128E MoE open-weight model with 1M token context, pretrained on ~22T tokens. Strong multimodal and multilingual capabilities for teams that need control, private deployment, and customization.

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

Why #2: gpt-oss-120b

OpenAI's 120B open-weight model for frontier-style reasoning with self-hosted deployment.

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

Why #3: Llama 3.3 70B Instruct

Meta's latest Llama 3.x dense model, 70B parameters with 128K context.

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