Instruction Following & Long Context

Artificial Analysis public leaderboard

Long Context Reasoning

Long-context reasoning results published by Artificial Analysis.

Published source

Metric

Accuracy

Results shown

516

Publisher

Artificial Analysis

Snapshot

Fetched 2026-09-08T22:57:37.545Z

516 of 516 models

Higher is better

Kimi K3 (max)

Claude Fable 5.1 (max with fallback)

Claude Fable 5.1 (medium with fallback)

GPT-5.5 (xhigh)

GPT-5.5 (high)

Gemini 3.8 Flash (medium)

GPT-5.6 Sol (max)

GPT-5.6 Luna (max)

Claude Fable 5.1 (high with fallback)

Muse Glimmer (high)

GPT-5.3 Codex (xhigh)

MiniMax-M3

GPT-5.5 (medium)

Muse Spark 1.3 (max)

Claude Fable 5.1 (xhigh with fallback)

GPT-5.6 Terra (max)

Muse Spark 1.3 (xhigh)

Gemini 3.7 Flash (medium)

Agnes 2.5 Pro Beta

GPT-5.2 (xhigh)

GPT-5.2 Codex (xhigh)

Claude Fable 5.1 (low with fallback)

Claude Fable 5 (with fallback)

GPT-5.6 Sol (xhigh)

Claude Sonnet 5 (max)

GPT-5.4 (xhigh)

Qwen3.8 27B (xhigh)

Gemini 3.1 Pro Preview

Claude Opus 5 (medium)

Nex-N2-Pro

GPT-5.6 Luna (xhigh)

GPT-5.6 Sol (high)

Gemini 3.7 Flash (high)

Claude Opus 5 (low)

DeepSeek V4 Flash Vision (max)

Gemini 3.8 Flash (high)

Kimi K2.6

Grok 4.6 (medium)

GPT-5.5 (low)

Grok 4.6 (xhigh)

GPT-6 Astra (max)

Qwen3.6 Max Preview

Grok 4.6 (low)

Gemini 3.8 Flash (low)

Claude Opus 5 (xhigh)

GPT-5.6 Sol (medium)

DeepSeek V4 Pro 0813 (max)

Qwen3.8 2.4T A95B

GPT-5.6 Luna (high)

Grok 4.6 (high)

GLM-5.3-Flash

GPT-6 Astra (xhigh)

GPT-5.1 (high)

Gemini 3.6 Flash

GPT-6 Astra (low)

K2 Horizon 375B A23B

Claude Sonnet 4.6 (max)

GPT-6 Astra (high)

MiMo-V2.5-Pro

Qwen3.8 27B (medium)

Qwen3.8-Flash-Next

GLM-5.3 (max)

GPT-6 Astra (medium)

DeepSeek V4 Flash 0731 (max)

Apodex 1.1

Kimi K3 (low)

Nemotron 3 Ultra

Grok 4.5 (high)

Kimi K2.7 Code

Claude Opus 5 (max)

Muse Spark 1.2 (xhigh)

Qwen3.7 Max

Claude Opus 5 (high)

GPT-5.6 Terra (xhigh)

Hy3

Gemini 3.7 Flash (low)

Claude Opus 4.7 (max)

MiniMax-M2.7

Qwen3.8 Max

Qwen3.6 Plus

GLM-5.2 (max)

GPT-5 (high)

GPT-5.6 Sol (low)

Claude Opus 4.6 (max)

Gemini 3 Flash

Kimi K2.5

Muse Spark

Qwen3.5 27B

GPT-5.6 Terra (high)

Claude Opus 4.8 (max)

Muse Spark 1.1 (xhigh)

Inkling

Qwen3.5 397B A17B

Claude Opus 4.5

Qwen3.8 27B (low)

Qwen3.6 27B

GPT-5.4 mini (xhigh)

GPT-5.4 (low)

Claude Sonnet 5 (xhigh)

Claude Sonnet 5 (high)

GPT-5.4 nano (xhigh)

Qwen3.5 122B A10B

Quasar 438B (max)

Gemini 3.5 Flash-Lite

GPT-5 (medium)

Motif 3

Claude 4.1 Opus

Gemini 3 Pro Preview (high)

Claude Opus 4.7 (Non-reasoning, high)

GLM-5

Inkling Small

Agnes 2.5 Pro Alpha

GPT-5.6 Luna (medium)

Grok 4.3 (medium)

MiMo-V2-Omni

Grok Build 0.1 0616

Motif 3 (Beta)

DeepSeek V4 Pro (max)

Gemini 3.5 Flash (medium)

Gemini 3.1 Flash-Lite

Qwen3 Max Thinking

Claude 4.5 Haiku

DeepSeek V4 Flash (max)

Solar Pro 4

Grok 4.1 Fast

GPT-5.6 Terra (medium)

Gemini 3 Pro Preview (low)

Grok 4.3 (low)

Grok 4 Fast

GLM-5.1

Step 3.7 Flash

Claude Sonnet 5 (medium)

Gemini 3.5 Flash

MiniMax-M2.5

MiMo-V2-Omni-0327

DeepSeek V3.2

MiMo-V2.5

Grok 4.3 (high)

Qwen3.7 Plus

Ling 3.0 Flash

KAT-Coder-Pro V2

GPT-5 mini (high)

Claude 4.5 Sonnet

DeepSeek V3.2 Exp

GPT-5 mini (medium)

Qwen3.5 35B A3B

DeepSeek V4 Flash (high)

Kimi K2 Thinking

Qwen3 235B A22B 2507

GLM-5-Turbo

Doubao Seed Code

GPT-5 Codex (high)

Solar Open2 250B

Qwen3.6 35B A3B

MiMo-V2-Flash (Feb 2026)

JT-4.1 Flash 236B A21B

GPT-5.6 Terra (low)

GLM-4.7

Gemini 2.5 Flash (Sep)

GPT-6 Astra (Non-reasoning)

Claude Opus 4.5 (Non-reasoning)

MiMo-V2-Flash

GLM 5V Turbo

DeepSeek V4 Pro (high)

GPT-5.2 (medium)

Claude 4 Sonnet

GPT-5.6 Luna (low)

Qwen3.5 9B

DeepSeek V3.2 Speciale

A.X-K2

GPT-5.5 Instant (June 2026)

Claude Sonnet 5 (Non-reasoning)

Ring-2.6-1T

Kimi K2.6 (Non-reasoning)

Gemma 4 31B

GPT-5.1 Codex (high)

DeepSeek V3.1 Terminus

Qwen3.8 27B

Mistral Medium 3.5

Claude Sonnet 4.6 (Non-reasoning, low)

Gemini 2.5 Pro

Grok 4.20 0309 v2

Claude Sonnet 4.6 (Non-reasoning, high)

GPT-4.1

MiMo-V2-Pro

Grok 4

Grok 4.20 0309

MiniMax-M2.1

Claude Sonnet 5 (low)

GPT-5.4 nano

Kimi K2.5 (Non-reasoning)

GPT-5.4 mini (medium)

Claude Opus 4.6 (Non-reasoning, high)

Qwen3.6 27B (Non-reasoning)

GPT-5.1 Codex mini (high)

GPT-5.5 Instant (May 2026)

JT-35B-Flash

Nemotron 3 Super

Gemma 4 26B A4B

Gemini 2.5 Flash

G9v3-39A5B

LongCat 2.0

GPT-5 (ChatGPT)

Gemini 2.5 Flash-Lite (Sep)

Hy3-preview

Nova 2.0 Pro Preview (low)

Qwen3.5 397B A17B (Non-reasoning)

Qwen3.6 35B A3B (Non-reasoning)

MiniMax-M2

Nova 2.0 Pro Preview (medium)

Qwen3.5 27B (Non-reasoning)

GPT-5.5 (Non-reasoning)

Qwen3.5 Omni Plus

Qwen3 VL 235B A22B (Reasoning)

Qwen3 Next 80B A3B (Reasoning)

Gemma 4 12B

Step 3.5 Flash 2603

Qwen3.5 4B

Qwen3.5 35B A3B (Non-reasoning)

GPT-5.6 Sol (Non-reasoning)

Qwen3 Max Thinking (Preview)

Qwen3.5 122B A10B (Non-reasoning)

Gemini 3.5 Flash (minimal)

Seed-OSS-36B-Instruct

Qwen3 30B A3B 2507

K-

K-EXAONE

o4-mini (high)

Nemotron 3.5 Lightning

Nova 2.0 Lite (high)

Ling 3.0 Tiny

K-

K-EXAONE 2.0

Gemini 2.5 Flash (Sep) (Non-reasoning)

Nova 2.0 Lite (medium)

Nova 2.0 Omni (medium)

MiniCPM5-2B

GPT-5.6 Terra (Non-reasoning)

GPT-5.4 (Non-reasoning)

Grok 3

DeepSeek R1 (Jan)

MiniMax M1 80k

Nova 2.0 Omni (low)

K2 Think V2

DeepSeek V3.1

GPT-4o (ChatGPT)

DeepSeek R1 0528

Gemini 2.5 Flash-Lite

Gemini 3 Flash (Non-reasoning)

Qwen3 VL 32B (Reasoning)

Grok 3 mini Reasoning (high)

EXAONE 4.5 33B

Ring-1T

Nova 2.0 Lite (low)

GLM-4.6

Claude 4.5 Sonnet (Non-reasoning)

GLM-5.1 (Non-reasoning)

K-

K-EXAONE (Non-reasoning)

Kimi K2

Kimi K2 0905

DeepSeek V4 Pro (Non-reasoning)

Magistral Medium 1.2

Grok Code Fast 1

GLM-4.5

Qwen3 Next 80B A3B

Command A+

Qwen3.5 Omni Flash

gpt-oss-120b (high)

Claude 3.7 Sonnet (Non-reasoning)

Qwen3 Max

Llama 4 Maverick

Step 3.5 Flash

Gemini 2.5 Flash (Non-reasoning)

Mistral Small 4

Gemini 2.5 Flash-Lite (Sep) (Non-reasoning)

Claude 4.5 Haiku (Non-reasoning)

GPT-4o (Nov)

Granite 4.2 30B

GLM-4.6V

DeepSeek V3.1 (Non-reasoning)

Qwen3 Coder Next

GLM-4.5-Air

Gemma 4 31B (Non-reasoning)

Qwen3.5 9B (Non-reasoning)

gpt-oss-120b (low)

Qwen3 Coder 480B

DeepSeek V3.2 (Non-reasoning)

GPT-5.2 (Non-reasoning)

DeepSeek V3.1 Terminus (Non-reasoning)

Granite 4.2 8B

GPT-5.1 (Non-reasoning)

GPT-5 nano (high)

DeepSeek V3.2 Exp (Non-reasoning)

GPT-4.1 mini

Claude 4 Sonnet (Non-reasoning)

Mercury 2

GPT-5 nano (medium)

GLM-5 (Non-reasoning)

o3-mini (high)

Qwen3 Max (Preview)

GPT-5.6 Luna (Non-reasoning)

Gemma 4 26B A4B (Non-reasoning)

GLM-5.2 (Non-reasoning)

Hy3-preview (Non-reasoning)

GLM-4.7-Flash

Ling-2.6-1T

MiMo-V2.5-Pro (Non-reasoning)

DeepSeek V4 Flash (Non-reasoning)

G9v3-3B

Solar Open 100B

NVIDIA Nemotron Nano 12B v2 VL

GPT-4o (Aug)

DeepSeek V3 0324

GLM-4.7 (Non-reasoning)

INTELLECT-3

Nemotron 3 Nano Omni 30B A3B

Nemotron Cascade 2 30B A3B

GPT-5 mini (minimal)

KAT-Coder-Pro V1

Nemotron 3 Nano

Celeris-1

Trinity Large Thinking

North Mini Code

Qwen3 4B 2507

GPT-5.4 mini (Non-reasoning)

Ling-1T

HyperNova 60B 2605 (high)

MiMo-V2-Flash (Non-reasoning)

Mistral Large 3

Gemma 4 12B (Non-reasoning)

K2-V2 (high)

Qwen3.5 4B (Non-reasoning)

gpt-oss-20b (high)

Nova Premier

Qwen3 235B 2507 (Non-reasoning)

Qwen3 VL 8B (Reasoning)

Qwen3 VL 235B A22B

Qwen3 Coder 30B A3B

Devstral 2

Grok 4.3 (Non-reasoning)

Solar Pro 3

Gemini 2.5 Flash-Lite (Non-reasoning)

Devstral Medium

Devstral Small (May)

Gemma 4 E4B

Grok 4.1 Fast (Non-reasoning)

Mistral Medium 3

Ling 2.6 Flash

Gemini 2.0 Flash

gpt-oss-20b (low)

Nova 2.0 Pro Preview (Non-reasoning)

GPT-5.4 nano (Non-reasoning)

DeepSeek V3 (Dec)

Mistral Small 4 (Non-reasoning)

LongCat Flash Lite

Kimi Linear 48B A3B Instruct

Devstral Small 2

K2-V2 (medium)

Llama 4 Scout

Claude 3 Haiku

Claude 3.5 Haiku

QwQ-32B

Ministral 3 14B

Qwen3 30B A3B 2507 (Non-reasoning)

GLM-4.6 (Non-reasoning)

Ministral 3 8B

Llama 3.1 405B

Grok 4.20 0309 (Non-reasoning)

Nova 2.0 Omni (Non-reasoning)

Llama Nemotron Super 49B v1.5 (Non-reasoning)

Granite 4.2 3B

Grok 4 Fast (Non-reasoning)

NVIDIA Nemotron Nano 9B V2 (Non-reasoning)

Gemma 4 E4B (Non-reasoning)

Granite 4.1 30B

Grok 4.20 0309 v2 (Non-reasoning)

Cogito v2.1

NVIDIA Nemotron Nano 9B V2

Mistral Small 3.1

Hermes 4 405B

Hermes 4 405B (Non-reasoning)

Tri-21B-think Preview

Nemotron 3 Nano 4B

Command A

Apriel-v1.5-15B-Thinker

Qwen3 VL 4B (Reasoning)

Ring-flash-2.0

Nova Pro

Qwen3.5 2B

NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)

Mistral Small 3.2

GPT-4.1 nano

GLM-4.7-Flash (Non-reasoning)

K2-V2 (low)

GPT-5 nano (minimal)

DiffusionGemma 26B A4B

Magistral Small 1.2

Jamba 1.7 Large

Nova Lite

Nova 2.0 Lite (Non-reasoning)

Devstral Small

Llama 3.3 Nemotron Super 49B

Llama 3.1 8B

Ling-flash-2.0

GLM-4.6V (Non-reasoning)

Ministral 3 3B

Qwen3 VL 8B

Gemma 4 E2B (Non-reasoning)

Gemma 4 E2B

Jamba 1.7 Mini

Tri-21B-Think

Llama 3.3 70B

Phi-4 Mini

DeepSeek R1 0528 Qwen3 8B

Motif-2-12.7B

Mi:dm K 2.5 Pro Preview

Qwen3.5 2B (Non-reasoning)

Qwen3 VL 4B

HyperCLOVA X SEED Think (32B)

Granite 4.1 8B

Nova Micro

JT-MINI

Llama 3.2 11B (Vision)

Llama 3.3 Nemotron Super 49B (Non-reasoning)

Granite 4.0 H Small

Qwen3 4B 2507 (Non-reasoning)

Falcon-H1R-7B

Nemotron 3 Nano (Non-reasoning)

DeepSeek R1 Distill Qwen 14B

DeepSeek R1 Distill Llama 70B

Hermes 4 70B

Qwen3.5 0.8B

Mi:dm K 2.5 Pro

DeepSeek R1 Distill Qwen 32B

Gemma 3 12B

Llama 3.1 Nemotron 70B

Qwen3.5 0.8B (Non-reasoning)

Gemma 3 27B

Granite 4.0 Micro

MiniCPM-V 4.6 1.3B

Granite 4.0 H 1B

Gemma 3 4B

Llama 3.2 1B

Jamba Reasoning 3B

MiniCPM5-1B

Granite 4.1 3B

Granite 4.0 1B

LFM2.5-2.6B

MiniCPM5-1B (Non-reasoning)

Hermes 4 70B (Non-reasoning)

Llama 3.2 3B

Mistral Large 2 (Jul)

DeepSeek R1 Distill Qwen 1.5B

GLM-4.5V (Non-reasoning)

Gemma 3 270M

Qwen3 Omni 30B A3B

DeepSeek R1 Distill Llama 8B

Qwen3 0.6B (Non-reasoning)

LFM2 1.2B

Magistral Small 1

Qwen3 4B

Mistral 7B

Qwen3 1.7B (Non-reasoning)

GLM-4.5V

Olmo 3 7B Think

Sarvam M

Olmo 3.1 32B Think

LFM2.5-VL-1.6B

Qwen3 30B

Molmo2-8B

Qwen3 235B (Non-reasoning)

Solar Pro 2 (Non-reasoning)

Qwen3 14B

Sarvam 105B (high)

LFM2.5-1.2B-Instruct

Apertus 70B Instruct

Qwen3 0.6B

OLMo 2 32B

Granite 4.0 350M

Qwen3 1.7B

LFM2.5-1.2B-Thinking

Magistral Medium 1

LFM2 2.6B

Sarvam 30B (high)

Nanbeige4.1-3B

Tiny Aya Global

Exaone 4.0 1.2B (Non-reasoning)

Llama 3 8B

Apertus 8B Instruct

Olmo 3 7B

Olmo 3.1 32B Instruct

Qwen3 8B

Granite 4.0 H 350M

Gemma 3n E4B

Gemma 3n E2B

Llama 3 70B

Step3 VL 10B

Molmo 7B-D

Qwen3 32B

Llama 3.1 Nemotron Nano 4B v1.1

Qwen3 8B (Non-reasoning)

Qwen3 Omni 30B A3B (Reasoning)

Qwen3 235B

Qwen3 32B (Non-reasoning)

Olmo 3 32B Think

Gemma 3 1B

Exaone 4.0 1.2B

Reka Flash 3

LFM2.5-8B-A1B

Mistral Small 3

Solar Pro 2

Qwen3 14B (Non-reasoning)

Phi-4

Qwen3 30B (Non-reasoning)

OLMo 2 7B

LFM2 24B A2B

List bars visualize relative performance. Plot bars use published 0–100 scores, shrink to a 1.5rem minimum, then scroll horizontally.

Scores are published snapshots and are only comparable within the same benchmark, version, metric, and evaluation configuration.