LLM AtlasLLM AtlasSearch models

Leaderboard

long context 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
long context

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
#1GPT-5.4OpenAI961,000,000
#2Claude Sonnet 4.6Anthropic951,000,000
#3Gemini 3.1 ProGoogle DeepMind951,048,576
#4GPT-5.2OpenAI951,000,000
#5Claude 3.7 SonnetAnthropic94200,000
#6Claude Sonnet 4.5Anthropic941,000,000
#7Gemini 2.5 ProGoogle DeepMind941,048,576
#8Gemini 3.0 ProGoogle DeepMind941,048,576
#9Gemini 3.1 FlashGoogle DeepMind941,048,576
#10GPT-4oOpenAI94128,000
#11GPT-5.3-CodexOpenAI941,000,000
#12Claude Opus 4.6Anthropic931,000,000
#13Claude Sonnet 4Anthropic931,000,000
#14Gemini 2.5 FlashGoogle DeepMind931,048,576
#15Gemini 2.5 Pro TTSGoogle DeepMind931,048,576
#16Gemini 3.0 FlashGoogle DeepMind931,048,576
#17GPT-5OpenAI931,000,000
#18GPT-5 miniOpenAI931,000,000
#19GPT-5.2 ProOpenAI931,000,000
#20GPT-5.3 InstantOpenAI931,000,000
#21GPT-5.4 ProOpenAI931,000,000
#22Claude Haiku 4.5Anthropic92200,000
#23Command R+ 2026Cohere92128,000
#24Gemini 2.0 FlashGoogle DeepMind921,048,576
#25GPT-4.1OpenAI921,048,576
#26Llama 4 MaverickMeta921,048,576
#27Mistral Large 25Mistral AI92128,000
#28Claude Opus 4.1Anthropic91200,000
#29Gemini 2.5 Flash LiveGoogle DeepMind911,048,576
#30Gemini 2.5 Flash Native Audio PreviewGoogle DeepMind911,048,576
#31Gemini 2.5 Flash-LiteGoogle DeepMind911,048,576
#32Gemini 3.1 Flash-LiteGoogle DeepMind911,048,576
#33o4-miniOpenAI91200,000
#34Claude Opus 4Anthropic90200,000
#35Gemini 1.5 ProGoogle DeepMind902,097,152
#36GPT-4.1 miniOpenAI901,048,576
#37GPT-5 nanoOpenAI901,000,000
#38o4-mini-deep-researchOpenAI90200,000
#39Sonar Reasoning ProPerplexity90200,000
#40Claude Haiku 3.5Anthropic89200,000
#41Gemini 1.5 FlashGoogle DeepMind891,048,576
#42Gemini 2.0 Flash-LiteGoogle DeepMind891,048,576
#43GPT-4o-miniOpenAI89128,000
#44morph-v3-fast-applyMorph89128,000
#45Nova ProAmazon Web Services89300,000
#46Sonar Deep ResearchPerplexity89200,000
#47Sonar ProPerplexity89200,000
#48warpgrep-v2Morph89128,000
#49Llama 3.3 70B InstructMeta88128,000
#50o1-miniOpenAI88128,000
#51o3OpenAI88200,000
#52SonarPerplexity88128,000
#53Claude Sonnet 3Anthropic87200,000
#54flash-compactMorph87200,000
#55gpt-oss-120bOpenAI87131,072
#56Nova LiteAmazon Web Services87300,000
#57o3-deep-researchOpenAI87200,000
#58Claude Haiku 3Anthropic86200,000
#59Claude Haiku 3Anthropic86200,000
#60Claude Haiku 3.5Anthropic86200,000
#61Claude Haiku 4.5Anthropic86200,000
#62Claude Opus 3Anthropic86200,000
#63CodestralMistral AI86256,000
#64Doubao-Seed-1.6ByteDance / Doubao86128,000
#65Doubao-Seed-1.6-FlashByteDance / Doubao86128,000
#66Doubao-Seed-2.0-CodeByteDance / Doubao86128,000
#67Doubao-Seed-CodeByteDance / Doubao86128,000
#68Gemini 1.5 Flash-8BGoogle DeepMind861,048,576
#69gpt-audioOpenAI86128,000
#70gpt-realtimeOpenAI86128,000
#71Llama 4 ScoutMeta8610,485,760
#72MiniMax-M1MiniMax86204,800
#73MiniMax-Text-01MiniMax86204,800
#74MiniMax-VL-01MiniMax86204,800
#75Step-3.5-FlashStepFun86131,072
#76Voxtral Mini TranscribeMistral AI86131,072
#77CogVideoXZ.AI85128,000
#78CogView 4Z.AI85128,000
#79DeepSeek-Coder-V2DeepSeek85128,000
#80DeepSeek-Math-V2DeepSeek85128,000
#81DeepSeek-R1DeepSeek85128,000
#82DeepSeek-R1-Distill-Llama-70BDeepSeek85128,000
#83DeepSeek-V2.5DeepSeek85128,000
#84DeepSeek-V3DeepSeek85128,000
#85DeepSeek-V3.1DeepSeek85128,000
#86DeepSeek-V3.1-BaseDeepSeek85128,000
#87DeepSeek-V3.2DeepSeek85128,000
#88DeepSeek-V3.2-ExpDeepSeek85128,000
#89Devstral 2 OpenMistral AI85128,000
#90Devstral Medium 1.0Mistral AI85128,000
#91Devstral Small 2Mistral AI85128,000
#92ERNIE 3.5 128KBaidu / ERNIE85128,000
#93ERNIE 4.0 Turbo 8KBaidu / ERNIE85128,000
#94ERNIE Functions 8KBaidu / ERNIE85128,000
#95ERNIE Speed 128KBaidu / ERNIE85128,000
#96GLM-4.5Z.AI85128,000
#97GLM-4.5VZ.AI85128,000
#98GLM-4.6Z.AI85128,000
#99GLM-4.6VZ.AI85128,000
#100GLM-4.7Z.AI85128,000
#101GLM-5Z.AI85128,000
#102GLM-ImageZ.AI85128,000
#103GLM-OCRZ.AI85128,000
#104Hunyuan CodeTencent / Hunyuan85128,000
#105Hunyuan LiteTencent / Hunyuan85128,000
#106Hunyuan StandardTencent / Hunyuan85128,000
#107Hunyuan T1Tencent / Hunyuan85256,000
#108Hunyuan T1 VisionTencent / Hunyuan85128,000
#109Hunyuan TurboSTencent / Hunyuan85128,000
#110Hunyuan TurboS LongText 128KTencent / Hunyuan85128,000
#111Jamba 3BAI21 Labs85256,000
#112Jamba LargeAI21 Labs85256,000
#113Jamba Large 1.6AI21 Labs85256,000
#114Jamba MiniAI21 Labs85256,000
#115Jamba Mini 1.6AI21 Labs85256,000
#116Jamba Mini 1.7AI21 Labs85256,000
#117Kimi K2Moonshot AI / Kimi85131,072
#118Kimi K2 ThinkingMoonshot AI / Kimi85256,000
#119Kimi K2 Turbo PreviewMoonshot AI / Kimi85256,000
#120Kimi K2.5Moonshot AI / Kimi85256,000
#121Llama 3.1 405B InstructMeta85128,000
#122Llama 3.1 70B InstructMeta85128,000
#123Magistral Medium 1.2Mistral AI85128,000
#124Magistral Small 1.2 OpenMistral AI85128,000
#125MiniMax-M2MiniMax85204,800
#126MiniMax-M2.1MiniMax85204,800
#127MiniMax-M2.1-highspeedMiniMax85204,800
#128MiniMax-M2.5MiniMax85204,800
#129MiniMax-M2.5-highspeedMiniMax85204,800
#130Ministral 3 14B OpenMistral AI85128,000
#131Ministral 3 3B OpenMistral AI85128,000
#132Ministral 3 8B OpenMistral AI85128,000
#133Mistral Large 3Mistral AI85128,000
#134Mistral Large 3 OpenMistral AI85128,000
#135Mistral Medium 3.1Mistral AI85128,000
#136Mistral Nemo 12BMistral AI85128,000
#137Mistral Small 3.1Mistral AI85128,000
#138Mistral Small 3.2 OpenMistral AI85128,000
#139Nova MicroAmazon Web Services85128,000
#140o1OpenAI85200,000
#141Phi-3-vision-128k-instructMicrosoft85128,000
#142Phi-3.5-mini-instructMicrosoft85131,072
#143Phi-3.5-MoE-instructMicrosoft85131,072
#144Phi-3.5-vision-instructMicrosoft85131,072
#145Phi-4-mini-flash-reasoningMicrosoft85131,072
#146Phi-4-mini-instructMicrosoft85131,072
#147Phi-4-multimodal-instructMicrosoft85131,072
#148Phi-4-reasoningMicrosoft85131,072
#149Phi-4-reasoning-plusMicrosoft85131,072
#150Phi-4-reasoning-vision-15BMicrosoft85131,072
#151Pixtral 12BMistral AI85131,072
#152Pixtral LargeMistral AI85131,072
#153Qwen2.5-1.5B-InstructAlibaba Qwen85131,072
#154Qwen2.5-14B-InstructAlibaba Qwen85131,072
#155Qwen2.5-32B-InstructAlibaba Qwen85131,072
#156Qwen2.5-3B-InstructAlibaba Qwen85131,072
#157Qwen2.5-72B-InstructAlibaba Qwen85131,072
#158Qwen2.5-7B-InstructAlibaba Qwen85131,072
#159Qwen2.5-MaxAlibaba Qwen85131,072
#160Qwen2.5-VL-72B-InstructAlibaba Qwen85131,072
#161Qwen2.5-VL-7B-InstructAlibaba Qwen85131,072
#162Qwen3-Coder-NextAlibaba Qwen85131,072
#163Qwen3.5-0.8BAlibaba Qwen85131,072
#164Qwen3.5-122B-A10BAlibaba Qwen85131,072
#165Qwen3.5-27BAlibaba Qwen85131,072
#166Qwen3.5-2BAlibaba Qwen85131,072
#167Qwen3.5-35B-A3BAlibaba Qwen85131,072
#168Qwen3.5-397B-A17BAlibaba Qwen85131,072
#169Qwen3.5-4BAlibaba Qwen85131,072
#170Qwen3.5-9BAlibaba Qwen85131,072
#171Vidu Q1Z.AI85128,000
#172Voxtral Mini OpenMistral AI85131,072
#173Voxtral Small OpenMistral AI85131,072
#174Claude Sonnet 3Anthropic84200,000
#175Claude Sonnet 4Anthropic841,000,000
#176Command ACohere84256,000
#177Command A ReasoningCohere84128,000
#178Command A TranslateCohere84128,000
#179Command A VisionCohere84128,000
#180Command R+Cohere84128,000
#181Command R7BCohere84128,000
#182Embed 4Cohere84128,000
#183Grok 3xAI84131,072
#184Grok 3 MinixAI84131,072
#185Grok 4xAI84256,000
#186Grok 4 Fast ReasoningxAI84131,072
#187grok-imagexAI84131,072
#188gpt-audio-miniOpenAI83128,000
#189gpt-oss-20bOpenAI83131,072
#190gpt-realtime-miniOpenAI83128,000
#191Llama 3.1 8B InstructMeta82128,000
#192Llama 3.2 90B Vision InstructMeta82128,000
#193Claude Opus 3Anthropic81200,000
#194Claude Opus 4Anthropic81200,000
#195Codestral EmbedMistral AI8132,768
#196ERNIE 4.5 Turbo 32KBaidu / ERNIE8132,768
#197Mistral EmbedMistral AI8132,768
#198Mistral ModerationMistral AI8132,768
#199Mistral OCR 2505Mistral AI8132,768
#200Llama 3.2 1B InstructMeta80128,000
#201Llama 3.2 3B InstructMeta80128,000
#202MiMo-VL-7BXiaomi80131,072
#203Step3-VL-10BStepFun80131,072
#204Llama 3.2 11B Vision InstructMeta79128,000
#205MiMo-Audio-7BXiaomi79131,072
#206GPT-4o mini TranscribeOpenAI78128,000
#207GPT-4o mini TTSOpenAI78128,000
#208GPT-4o TranscribeOpenAI78128,000
#209LFM2-24B-A2BLiquid AI7832,768
#210Step-Audio-R1.1StepFun78131,072
#211DeepSeek-OCRDeepSeek7716,384
#212DeepSeek-OCR-2DeepSeek7716,384
#213DeepSeek-VL2-SmallDeepSeek7716,384
#214Janus-Pro-7BDeepSeek7716,384
#215Phi-4Microsoft7716,384
#216image-01MiniMax768,192
#217image-01-liveMiniMax768,192
#218Llama Guard 4 12BMeta76131,072
#219MiniMax-Speech-02MiniMax768,192
#220music-2.0MiniMax768,192
#221LFM2-8B-A1BLiquid AI7532,768
#222LFM2.5-1.2B-InstructLiquid AI75131,072
#223LFM2.5-1.2B-ThinkingLiquid AI75131,072
#224GPT Image 1OpenAI7432,768
#225Llama Guard 3 11B VisionMeta74131,072
#226chatgpt-image-latestOpenAI7332,768
#227gpt-image-1-miniOpenAI7332,768
#228Code Llama 70B InstructMeta718,192
#229Meta Llama 3 70B InstructMeta718,192
#230phi-1Microsoft714,096
#231phi-1_5Microsoft714,096
#232phi-2Microsoft714,096
#233Phi-3-medium-4k-instructMicrosoft714,096
#234Phi-3-mini-4k-instructMicrosoft714,096
#235Phi-tiny-MoE-instructMicrosoft714,096
#236Code Llama 34B InstructMeta698,192
#237LFM2-2.6BLiquid AI6932,768
#238Meta Llama 3 8B InstructMeta698,192
#239pplx-embed-v1-4bPerplexity638,192
#240pplx-embed-v1-0.6bPerplexity628,192
#241FLUX 1.1 ProBlack Forest Labs50512
#242FLUX 1 ProBlack Forest Labs49512
#243FLUX 1.1 Pro UltraBlack Forest Labs49512
#244Prompt Guard 86MMeta47512
#245NextStep-1.1StepFun43512

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.