Knowledge

Artificial Analysis public leaderboard

Humanity's Last Exam

Broad expert-level reasoning and knowledge evaluation published by Artificial Analysis.

Published source

Metric

Accuracy

Results shown

607

Publisher

Artificial Analysis

Snapshot

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

607 of 607 models

Higher is better

Claude Fable 5.1 (max with fallback)

Claude Fable 5.1 (xhigh with fallback)

Claude Fable 5.1 (high with fallback)

Claude Fable 5 (with fallback)

Claude Opus 5 (max)

GPT-6 Astra (max)

GPT-6 Astra (xhigh)

Claude Opus 5 (xhigh)

Claude Fable 5.1 (medium with fallback)

GPT-6 Astra (high)

Claude Opus 5 (high)

GPT-6 Astra (medium)

Claude Opus 5 (medium)

GPT-5.6 Sol (max)

GPT-6 Astra (low)

Claude Fable 5.1 (low with fallback)

Muse Spark 1.3 (max)

Claude Opus 4.8 (max)

Gemini 3.7 Flash (high)

Gemini 3.8 Flash (high)

Muse Spark 1.3 (xhigh)

GPT-5.6 Sol (xhigh)

Gemini 3.1 Pro Preview

Kimi K3 (max)

Muse Spark 1.1 (xhigh)

GPT-5.6 Sol (high)

GPT-5.5 (xhigh)

Muse Spark 1.2 (xhigh)

GPT-5.5 (high)

Grok 4.6 (xhigh)

GPT-5.4 (xhigh)

Claude Opus 5 (low)

Qwen3.8 Max

GPT-5.6 Terra (max)

Grok 4.6 (high)

Gemini 3.5 Flash

Grok 4.5 (high)

GPT-5.3 Codex (xhigh)

Qwen3.8 2.4T A95B

GPT-5.5 (medium)

Claude Opus 4.7 (max)

GLM-5.3 (max)

GPT-5.6 Sol (medium)

Grok 4.6 (medium)

Gemini 3.8 Flash (medium)

GPT-5.6 Terra (xhigh)

Gemini 3.5 Flash (medium)

Claude Sonnet 5 (max)

GLM-5.2 (max)

DeepSeek V4 Pro 0813 (max)

Gemini 3.6 Flash

Muse Spark

Qwen3.7 Max

Motif 3 (Beta)

Claude Opus 4.6 (max)

GLM-5.3-Flash

Gemini 3 Pro Preview (high)

GPT-5.6 Luna (max)

GPT-5.6 Sol (low)

Claude Sonnet 5 (xhigh)

MiniMax-M3

Gemini 3.7 Flash (medium)

DeepSeek V4 Flash 0731 (max)

GPT-5.6 Terra (high)

Grok Build 0.1 0616

Qwen3.8-Flash-Next

GPT-5.2 (xhigh)

DeepSeek V4 Pro (max)

Agnes 2.5 Pro Beta

Kimi K2.6

Grok 4.3 (high)

GPT-6 Astra (Non-reasoning)

Gemini 3.8 Flash (low)

GPT-5.6 Luna (xhigh)

Motif 3

Gemini 3 Flash

GPT-5.2 Codex (xhigh)

Claude Sonnet 5 (high)

MiMo-V2.5-Pro

Qwen3.7 Plus

DeepSeek V4 Pro (high)

Gemini 3.7 Flash (low)

Kimi K2.7 Code

DeepSeek V4 Flash (max)

Grok 4.20 0309 v2

DeepSeek V4 Flash Vision (max)

Apodex 1.1

Qwen3.8 27B (xhigh)

Nex-N2-Pro

LongCat 2.0

KAT-Coder-Pro V1

Agnes 2.5 Pro Alpha

Claude Sonnet 4.6 (max)

Hy3

GPT-5.6 Luna (high)

Claude Opus 4.7 (Non-reasoning, high)

Inkling Small

GPT-5.6 Terra (medium)

GPT-5.5 (low)

Grok 4.20 0309

K2 Horizon 375B A23B

Inkling

GPT-5.4 (low)

Qwen3.6 Max Preview

Kimi K2.5

MiMo-V2-Pro

DeepSeek V4 Flash (high)

Claude Opus 4.5

GLM-5.1

Grok 4.3 (medium)

Claude Sonnet 5 (medium)

MiniMax-M2.7

A.X-K2

Gemini 3 Pro Preview (low)

GLM-5

Solar Pro 4

GPT-5.6 Terra (low)

Qwen3.5 397B A17B

DeepSeek V3.2 Speciale

Solar Open2 250B

GPT-5 (high)

GPT-5.1 (high)

Nemotron 3 Ultra

GPT-5.4 nano (xhigh)

GPT-5.4 mini (xhigh)

Qwen3 Max Thinking

Grok 4.20 0309 v2 (Non-reasoning)

GLM-5.1 (Non-reasoning)

GPT-5 Codex (high)

Qwen3.6 Plus

GLM-5-Turbo

Hy3-preview

Grok 4.6 (low)

GLM-4.7

MiMo-V2.5

Grok 4

GPT-5.2 (medium)

GPT-5.6 Luna (medium)

GPT-5.1 Codex (high)

GPT-5 (medium)

Qwen3.5 122B A10B

Kimi K3 (low)

DeepSeek V3.2

Grok 4.20 0309 (Non-reasoning)

Step 3.5 Flash 2603

Gemini 3.5 Flash (minimal)

Qwen3.5 27B

Kimi K2 Thinking

Ling 3.0 Flash

Gemma 4 31B

MiniMax-M2.1

Qwen3.6 27B

MiMo-V2-Flash

MiMo-V2-Omni-0327

Gemini 2.5 Pro

Qwen3.6 35B A3B

MiMo-V2-Flash (Feb 2026)

MiMo-V2-Omni

Muse Glimmer (high)

Claude Sonnet 5 (low)

GPT-5.5 Instant (May 2026)

Ring-2.6-1T

GPT-5 mini (high)

Step 3.7 Flash

Step 3.5 Flash

Qwen3.5 35B A3B

Nemotron 3 Super

MiniMax-M2.5

Gemini 2.5 Pro (May)

GPT-5.5 Instant (June 2026)

GPT-5.6 Luna (low)

Qwen3.5 397B A17B (Non-reasoning)

gpt-oss-120b (high)

Kimi K2.6 (Non-reasoning)

GPT-5 (low)

Grok 4.1 Fast

Gemma 4 26B A4B

Claude Opus 4.6 (Non-reasoning, high)

Grok 4 Fast

Claude Sonnet 5 (Non-reasoning)

Gemini 3.5 Flash-Lite

Quasar 438B (max)

K-

K-EXAONE 2.0

GPT-5.4 mini (medium)

GPT-5.1 Codex mini (high)

Grok 4.3 (low)

Gemini 2.5 Pro (Mar)

JT-4.1 Flash 236B A21B

Claude 4.5 Sonnet

G9v3-39A5B

Gemini 3.1 Flash-Lite

Mercury 2

GLM 5V Turbo

GPT-5.6 Sol (Non-reasoning)

HyperNova 60B 2605 (high)

o4-mini (high)

DeepSeek V3.1 Terminus

KAT-Coder-Pro V2

Qwen3.5 122B A10B (Non-reasoning)

GPT-5.4 nano

GPT-5 mini (medium)

Qwen3 235B A22B 2507

DeepSeek R1 0528

Trinity Large Thinking

Gemma 4 12B

Qwen3.6 27B (Non-reasoning)

Gemini 3 Flash (Non-reasoning)

Qwen3.5 9B

Qwen3.5 Omni Plus

DeepSeek V3.2 Exp

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

GLM-4.6

DeepSeek V3.1

Doubao Seed Code

Qwen3.8 27B (medium)

Qwen3.8 27B (low)

Qwen3.5 27B (Non-reasoning)

K-

K-EXAONE

Qwen3.6 35B A3B (Non-reasoning)

Gemini 2.5 Flash (Sep)

Mistral Medium 3.5

GPT-5.5 (Non-reasoning)

MiniMax-M2

Qwen3.5 35B A3B (Non-reasoning)

Claude Sonnet 4.6 (Non-reasoning, high)

ERNIE 5.0 Thinking Preview

Claude Opus 4.5 (Non-reasoning)

Kimi K2.5 (Non-reasoning)

INTELLECT-3

GLM-4.5

EXAONE 4.5 33B

Qwen3 Max Thinking (Preview)

Qwen3 Next 80B A3B (Reasoning)

Claude 4.1 Opus

Claude 4 Opus

Gemini 2.5 Flash

Qwen3.8 27B

Apriel-v1.5-15B-Thinker

Gemini 2.5 Flash (Apr)

o3-mini (high)

Nemotron Cascade 2 30B A3B

Cogito v2.1

Command A+

Qwen3 Max

Qwen3 VL 235B A22B (Reasoning)

Gemma 4 31B (Non-reasoning)

Nova 2.0 Lite (high)

Gemma 4 26B A4B (Non-reasoning)

EXAONE 4.0 32B

Nemotron 3 Nano

GPT-5.6 Terra (Non-reasoning)

GPT-5.4 (Non-reasoning)

DeepSeek V3.2 (Non-reasoning)

Claude Sonnet 4.6 (Non-reasoning, low)

Granite 4.2 30B

Qwen3 235B 2507 (Non-reasoning)

Ring-1T

North Mini Code

Sarvam 105B (high)

Grok 3 mini Reasoning (high)

gpt-oss-20b (high)

Falcon-H1R-7B

Qwen3 235B

Nanbeige4.1-3B

Hermes 4 405B

Step3 VL 10B

DiffusionGemma 26B A4B

Apriel-v1.6-15B-Thinker

Claude 4 Sonnet

Nemotron 3.5 Lightning

K2-V2 (high)

Solar Open 100B

Claude 4.5 Haiku

Qwen3 30B A3B 2507

Magistral Medium 1.2

Solar Pro 3

Qwen3 Max (Preview)

Qwen3 VL 32B (Reasoning)

Qwen3 Coder Next

K2 Think V2

Qwen3.5 4B

Mistral Small 4

Seed-OSS-36B-Instruct

Magistral Medium 1

GLM-5.2 (Non-reasoning)

Granite 4.2 8B

Claude 3.7 Sonnet

Ring-flash-2.0

GLM-4.6V

GPT-5 nano (high)

Nova 2.0 Pro Preview (medium)

Qwen3.5 9B (Non-reasoning)

Ling 3.0 Tiny

Mi:dm K 2.5 Pro Preview

DeepSeek V3.2 Exp (Non-reasoning)

Nova 2.0 Lite (medium)

MiniCPM5-2B

Qwen3 VL 30B A3B (Reasoning)

MiniMax M1 80k

Motif-2-12.7B

Hermes 4 70B

Ling-2.6-1T

Gemini 2.5 Flash (Sep) (Non-reasoning)

GPT-5 nano (medium)

DeepSeek V3.1 Terminus (Non-reasoning)

MiMo-V2-Flash (Non-reasoning)

DeepSeek R1 (Jan)

DeepSeek V4 Pro (Non-reasoning)

Mi:dm K 2.5 Pro

GPT-5.2 (Non-reasoning)

Grok Code Fast 1

Qwen3.5 4B (Non-reasoning)

o3-mini

MiniMax M1 40k

DeepSeek V4 Flash (Non-reasoning)

Qwen3 Next 80B A3B

GLM-5 (Non-reasoning)

GLM-4.7-Flash

Qwen3.5 Omni Flash

Magistral Small 1

Sarvam 30B (high)

Qwen3 Omni 30B A3B (Reasoning)

Kimi K2

Qwen3 32B

Llama Nemotron Ultra

Solar Pro 2

QwQ-32B

Ling-1T

Llama Nemotron Super 49B v1.5

Claude 4.5 Sonnet (Non-reasoning)

GPT-5.6 Luna (Non-reasoning)

GLM-4.5-Air

Nova 2.0 Omni (medium)

Gemini 2.5 Flash-Lite (Sep)

Hy3-preview (Non-reasoning)

Qwen3 30B A3B 2507 (Non-reasoning)

LFM2.5-8B-A1B

Qwen3 VL 32B

Celeris-1

Gemini 2.5 Flash-Lite

Grok 4.3 (Non-reasoning)

LFM2.5-1.2B-Instruct

DeepSeek V3.1 (Non-reasoning)

Granite 4.2 3B

Qwen3 VL 235B A22B

GPT-5 (ChatGPT)

Sonar Pro

MiniCPM5-1B

JT-35B-Flash

JT-MINI

Magistral Small 1.2

Granite 4.0 H 350M

Olmo 3 32B Think

Kimi K2 0905

GLM-4.7 (Non-reasoning)

Gemini 2.0 Flash Thinking exp. (Jan)

Gemma 4 12B (Non-reasoning)

GLM-4.5V

Qwen3 VL 30B A3B

Olmo 3.1 32B Think

Ling 2.6 Flash

LFM2.5-1.2B-Thinking

Ling-flash-2.0

LFM2.5-2.6B

Gemini 2.0 Pro Experimental

Qwen3 4B 2507

Claude 4 Opus (Non-reasoning)

Qwen3 30B

Solar Pro 2 (Preview)

GPT-5 (minimal)

Olmo 3 7B Think

Exaone 4.0 1.2B

Tri-21B-Think

DeepSeek R1 0528 Qwen3 8B

GPT-5.4 mini (Non-reasoning)

gpt-oss-120b (low)

LongCat Flash Lite

Tri-21B-think Preview

Olmo 3 7B

Exaone 4.0 1.2B (Non-reasoning)

K-

K-EXAONE (Non-reasoning)

Qwen3 0.6B

Llama 3.3 Nemotron Super 49B

LFM2 1.2B

Llama 3.2 11B (Vision)

Granite 4.0 350M

NVIDIA Nemotron Nano 12B v2 VL

HyperCLOVA X SEED Think (32B)

Apertus 70B Instruct

LFM2 2.6B

GLM-4.6 (Non-reasoning)

Llama 3.2 1B

Ministral 3 3B

OLMo 2 7B

DeepSeek Coder V2 Lite

Llama 3.2 3B

Llama 2 Chat 7B

Gemma 3 1B

Qwen3 1.7B (Non-reasoning)

Gemma 3 4B

Llama 3.1 8B

gpt-oss-20b (low)

GPT-5.1 (Non-reasoning)

Tiny Aya Global

Llama 2 Chat 70B

Llama 3.1 Nemotron Nano 4B v1.1

Nova 2.0 Pro Preview (low)

Molmo 7B-D

Jamba 1.5 Mini

DeepSeek R1 Distill Llama 70B

LFM2.5-VL-1.6B

Qwen3.5 0.8B (Non-reasoning)

Grok 4.1 Fast (Non-reasoning)

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

Llama 3 8B

Ling-mini-2.0

MiniCPM-V 4.6 1.3B

GPT-5 mini (minimal)

GPT-4.1 mini

Granite 4.0 Micro

EXAONE 4.0 32B (Non-reasoning)

Olmo 3.1 32B Instruct

Granite 4.0 H 1B

Phi-4 Multimodal

Phi-3 Mini

Qwen3.5 2B (Non-reasoning)

Apertus 8B Instruct

GLM-4.7-Flash (Non-reasoning)

Qwen3 0.6B (Non-reasoning)

LFM 40B

LFM2 8B A1B

Llama 4 Maverick

Qwen2.5 Coder 7B

Sonar

Nemotron 3 Nano 4B

NVIDIA Nemotron Nano 9B V2

Gemma 4 E4B (Non-reasoning)

Nemotron 3 Nano Omni 30B A3B

Granite 4.0 1B

Gemma 3n E4B (May)

Command-R (Mar)

OpenChat 3.5

Gemma 4 E2B

Llama 2 Chat 13B

DeepSeek V3 0324

Gemini 2.5 Flash (Non-reasoning)

Gemma 4 E2B (Non-reasoning)

Gemini 1.5 Flash-8B

Mistral Medium 3.1

Mixtral 8x7B

Gemini 1.5 Pro (Sep)

Qwen3 1.7B

Qwen3 Omni 30B A3B

Nova Micro

Ministral 3 14B

Command-R+ (Apr)

Qwen3 30B (Non-reasoning)

Nemotron 3 Nano (Non-reasoning)

NVIDIA Nemotron Nano 9B V2 (Non-reasoning)

Qwen3 VL 4B (Reasoning)

DeepSeek R1 Distill Qwen 32B

Mistral 7B

Qwen3 Coder 480B

Grok 4 Fast (Non-reasoning)

MiniCPM5-1B (Non-reasoning)

Qwen3 14B

Llama 3 70B

G9v3-3B

Gemma 3n E4B

Qwen3 4B 2507 (Non-reasoning)

Llama 3.2 90B (Vision)

Llama 3.1 70B

Phi-4 Mini

Grok Beta

Jamba 1.7 Mini

Reka Flash 3

Qwen3 4B

Gemma 3 27B

Gemini 1.5 Flash (May)

K2-V2 (medium)

Jamba 1.6 Mini

Mistral Small 3.1

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

Mistral Small 3.2

Mistral Saba

Nova Lite

Molmo2-8B

Ministral 3 8B

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

Claude 4 Sonnet (Non-reasoning)

DeepHermes 3 - Llama-3.1 8B

Gemini 2.0 Flash

Mistral Small (Sep)

Claude 4.5 Haiku (Non-reasoning)

Hermes 4 405B (Non-reasoning)

LFM2 24B A2B

Gemma 3 12B

GPT-4o mini

Qwen3 235B (Non-reasoning)

Llama 3.1 Nemotron 70B

Gemini 1.0 Pro

GPT-4.1

Granite 3.3 8B

Mistral Large 3

Gemma 3n E2B

Nova Premier

Claude 3.7 Sonnet (Non-reasoning)

DeepSeek R1 Distill Qwen 14B

Qwen3 32B (Non-reasoning)

Granite 4.1 30B

Grok 3

Qwen2.5 Turbo

Mistral Small (Feb)

Gemini 2.0 Flash-Lite (Preview)

Qwen3 14B (Non-reasoning)

Claude 3 Haiku

GPT-5.4 nano (Non-reasoning)

Jamba 1.5 Large

Gemini 2.0 Flash (exp)

Mistral Medium 3

Command A

Nova 2.0 Lite (low)

Devstral Small (May)

Mixtral 8x22B

Qwen2.5 Instruct 32B

Hermes 3 - Llama-3.1 70B

GPT-5 nano (minimal)

Llama 3.1 405B

GPT-4o (Mar)

Nova 2.0 Pro Preview (Non-reasoning)

QwQ 32B-Preview

Qwen3 8B

Claude 2.1

Nova 2.0 Omni (Non-reasoning)

Qwen3 Coder 30B A3B

Qwen3 VL 8B (Reasoning)

Qwen2.5 Max

Llama 3.3 Nemotron Super 49B (Non-reasoning)

Gemini 2.5 Flash (Apr) (Non-reasoning)

Mistral Small 4 (Non-reasoning)

Granite 4.0 H Small

Granite 4.1 8B

Llama 4 Scout

DeepHermes 3 - Mistral 24B

Mistral Small 3

Phi-4

GPT-4.1 nano

Devstral Small

Devstral Medium

Gemma 4 E4B

Jamba Reasoning 3B

Solar Pro 2 (Non-reasoning)

Claude 3.5 Sonnet (Oct)

Gemini 2.5 Flash-Lite (Non-reasoning)

GLM-4.6V (Non-reasoning)

Solar Pro 2 (Preview) (Non-reasoning)

Jamba 1.7 Large

Qwen2 72B

Claude 3.5 Haiku

Claude 3 Sonnet

Gemma 3 270M

OLMo 2 32B

K2-V2 (low)

Qwen3 VL 4B

Jamba 1.6 Large

o1-mini

Devstral 2

Hermes 4 70B (Non-reasoning)

Llama 3.3 70B

Qwen2.5 72B

Claude Instant

Qwen2.5 Coder 32B

Mistral Medium

Mistral Large (Feb)

Gemini 1.5 Pro (May)

GLM-4.5V (Non-reasoning)

Devstral Small 2

Granite 4.1 3B

Gemini 2.0 Flash-Lite (Feb)

Mistral Large 2 (Nov)

ERNIE 4.5 300B A47B

Qwen3 4B (Non-reasoning)

Tulu3 405B

Claude 3.5 Sonnet (June)

Nova Pro

Gemini 1.5 Flash (Sep)

GPT-4 Turbo

Sarvam M

DeepSeek R1 Distill Qwen 1.5B

Grok 2

Mistral Large 2 (Jul)

DBRX

DeepSeek V3 (Dec)

Nova 2.0 Lite (Non-reasoning)

Pixtral Large

Claude 3 Opus

GPT-4o (ChatGPT)

Qwen3 VL 8B

Qwen3.5 2B

Kimi Linear 48B A3B Instruct

GPT-4o (Nov)

GPT-4o (Aug)

Qwen3 8B (Non-reasoning)

GPT-4o (May)

Qwen3.5 0.8B

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.