Knowledge & reasoning
Global-MMLU-Lite leaderboard
As of 2026-10-07, the highest measured score on Global-MMLU-Lite is 93.2% by Gemini 3.1 Pro Preview.
Measured scores: artificial-analysis.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | Gemini 3.1 Pro Preview | 93.2% | measured |
| 2 | Gemini 3 Flash (non-reasoning) | 92.7% | measured |
| 3 | Gemini 3 Pro Preview (high) | 92.2% | measured |
| 4 | Claude Opus 4.6 (max) | 92.2% | measured |
| 5 | Gemini 3 Flash | 91.4% | measured |
All leaderboards
Agentic · terminal
Agentic · tools
- AA-AnalystAgent
- APEX-Agents
- AutomationBench
- EnterpriseOps-Gym
- Harvey LAB
- ITBench SRE
- τ³-Bench Banking
- τ²-Bench Telecom