Skip to content

Models

Comparative table of the 111 frontier models × 31 benchmarks. Click any header to sort. Heatmap per column (red = worst of filtered set, green = best). Frontier Index ranking on home.

Filter by tier
43 models · 31 benchmarks
Categories:(all — click to filter)
Model MMLUMMLU-ProGPQA-DiamondBBHARC-AGI-2Humanitys-Last-ExamMMMUHumanEvalMBPP+SWE-bench-VerifiedSWE-bench-ProCyberGymLiveCodeBenchAider-polyglotTerminal-Bench-HardTerminal-Bench-2MATH-500AIME-2024AIME-2025GSM8KFrontierMathSimpleQAIFEvalArena-HardMGSMTAU-benchOSWorldBrowseCompGDPvalArena-ELOLiveBench
GPT-5.5
OpenAI · 2026-04
93.685.058.682.735.478.784.484.9
GPT-5.5 Pro
OpenAI · 2026-04
90.1
GPT-5.4
OpenAI · 2026-03
92.880.057.775.083.0
Claude Opus 4.7
Anthropic · 2026-04
91.594.254.787.664.369.478.079.3
Claude Mythos Preview
Anthropic · 2026-04
94.656.893.977.883.182.079.686.9
Claude Sonnet 4.6
Anthropic · 2026-02
74.160.479.697.872.5
Gemini 3 Deep Think
Google DeepMind · 2026-02
84.648.4
Gemini 3.1 Pro
Google DeepMind · 2026-02
91.494.377.144.480.668.591.272.190.085.9
Gemma 4 (31B dense)open
Google DeepMind · 2026-04
85.284.380.089.2
Grok 4.3
xAI · 2026-04
Grok 4.20
xAI · 2026-03
70.8
Muse Spark
Meta · 2026-04
58.0
Mistral Medium 3.5
Mistral AI · 2026-04
77.6
Mistral Small 4open
Mistral AI · 2026-03
78.071.2
Command Aopen
Cohere · 2025-03
85.550.880.090.951.7
Command A Reasoningopen
Cohere · 2025-08
Reka Flash 3.1open
Reka · 2025-07
66.953.5
Jamba2 Mini
AI21 Labs · 2026-01
DeepSeek V4 Proopen
DeepSeek · 2026-04
87.590.137.776.880.655.493.587.592.6
DeepSeek V4 Flashopen
DeepSeek · 2026-04
Qwen3.6 Max Preview
Alibaba · 2026-04
Qwen3.6-27B
Alibaba · 2026-04
77.253.583.959.3
Qwen3.6-Plusopen
Alibaba · 2026-04
86.078.861.6
GLM-5
Zhipu AI · 2026-02
86.050.477.892.7
GLM-5.1open
Zhipu AI · 2026-03
77.858.4
ERNIE 5.1 Preview
Baidu · 2026-04
ERNIE 5.0
Baidu · 2026-01
Doubao Seed 2.0 Pro
ByteDance · 2026-02
87.088.985.476.587.854.298.387.4
Yi-Lightning
01.AI · 2024-10
76.050.983.576.4
MiMo V2.5 Pro
Xiaomi · 2026-04
MiMo V2.5
Xiaomi · 2026-04
76.056.1
MiniMax M2.7open
MiniMax · 2026-03
78.056.257.0
Step 3.5 Flashopen
StepFun · 2026-02
74.486.451.097.331.6
Nemotron 3 Nano Omni
Nvidia · 2026-04
Nemotron 3 Superopen
Nvidia · 2026-03
86.083.379.417.479.460.581.290.273.9
AFM Server
Apple · 2025-07
80.089.174.6
AFM On-Device
Apple · 2025-07
67.985.174.9
Amazon Nova 2 Omni
Amazon · 2025-12
Nova 2 Pro
Amazon · 2025-12
Samsung Gauss 2.3
Samsung · 2025-09
Kimi K2.6open
Moonshot AI · 2026-04
90.534.780.258.689.666.773.183.2
EXAONE 4.5 33Bopen
LG AI Research · 2026-04
83.380.581.492.9
K-EXAONE 236B-A23Bopen
LG AI Research · 2026-01
83.879.113.649.480.792.8