← Gemini 3.1 Pro · API · High thinking
Evidenceacceptedindependent_lab
MMMU-Pro
Artificial Analysis — MMMU-Pro (Gemini 3.1 Pro Preview)
ev_2026_07_gem31pro_aa_mmmu
Result
82%
Confidence A
- Tested
- 2026-07-11
- Runs
- 1
Configuration
How it was measured
- Exact model
- gemini-3.1-pro-preview
- Reasoning
- high
- Agent harness
- none
- Tool access
- none (no-tools academic multimodal)
- Prompting
- Artificial Analysis MMMU-Pro protocol
PRIMARY UI / multimodal — independent Artificial Analysis MMMU-Pro. AA and secondary press report Gemini 3.1 Pro Preview at ~82% (second to Gemini 3.5 Flash high ~84% on later boards). Unlocks Gemini slot_ui_visual.
Caveats
Known limitations
- Academic multimodal score — not a pure Figma/UI craft preference test.
- Float transcribed from AA leaderboard coverage (chart + secondary AA-sourced reports).
- High/thinking config — not Flash.
Weighting
Evidence weight factors
The engine weights each evidence row by these factors (0–1) when fusing scores — higher is more trusted.
relevance0.95
quality1.00
recency1.00
health0.85
comparability0.90
reliability0.85
Benchmark
MMMU-Pro
Multimodal understanding and reasoning: college-level questions that require interpreting images (diagrams, charts, photos, figures) alongside text. MMMU-Pro is a harder, cleaner follow-on to MMMU that reduces shortcut strategies. Primary ModelQuest signal for visual / UI-relevant understanding when formal design rubrics are unavailable.
Benchmark source →