Registry
1573models
206 benchmarks17 categories944 scores
Updated 1w ago
Beta version: *Information might not be fully accurate. Please report any discrepancies.
Registry / Live Benchmarks
BetaGlobal LLM Leaderboard
Tracking performance, provenance, and variants across verified foundation models.
New
Google DeepMindGemini 3.1 Pro
Feb 19 · 14 benchmarks
Performance by Domain
About DomainsIntelligence
Top: o3-pro
85pts· n=35
Knowledge & Communication
Top: Gemini 3.1 Pro
93pts· n=6
Coding
Top: GPT-5.1
98pts· n=10
Math
Top: GPT-5.2 Pro
100pts· n=11
Agents & Tools
Top: Gemini 2.0 Flash
94pts· n=48
Vision & Video
Top: GPT-5.2 Pro
89pts· n=88
Long Context
Top: Gemini 3.1 Pro
85pts· n=4
Factuality
Top: Grok-4.1-Fast
97pts· n=4