Vals Science Index ↗27.38 %
Gemini 3.1 Pro Preview (02/26) · Vals Science Index · Accuracy ↗
Google / MODEL EVIDENCE
Gemini 3.1 Pro Preview (02/26)
32 benchmarks · 35 observations · 35 sources
Scores at a glance
32 benchmarks · large dot: Gemini 3.1 Pro Preview (02/26); small dots: closest models where reported
Vals Multimodal Index ↗56.07 %
Gemini 3.1 Pro Preview (02/26) · Vals Multimodal Index · Accuracy ↗
Legal Research Bench ↗20.67 %
Gemini 3.1 Pro Preview (02/26) · Legal Research Bench · All-Pass ↗
LegalBench ↗87.4 %
Gemini 3.1 Pro Preview (02/26) · LegalBench · Accuracy ↗
Gemini 3.1 Pro Preview (02/26) · Harvey's Legal Agent Benchmark v1 · Accuracy ↗
CaseLaw v2 ↗64.84 %
Gemini 3.1 Pro Preview (02/26) · CaseLaw · Accuracy ↗
Tax Agent Bench ↗9.29 %
Gemini 3.1 Pro Preview (02/26) · Tax Agent Bench · Accuracy ↗
Vals Finance Agent v2 ↗42.98 %
Gemini 3.1 Pro Preview (02/26) · Finance Agent v2 · Accuracy ↗
Gemini 3.1 Pro Preview (02/26) · Excel Modeling Benchmark v1 · Accuracy ↗
CorpFin v2 ↗64.49 %
Gemini 3.1 Pro Preview (02/26) · CorpFin · Accuracy ↗
Gemini 3.1 Pro Preview (02/26) · MortgageTax · Accuracy ↗
1 more reported result
72.88 % · Gemini 3.1 Pro Preview (02/26) · TaxEval · Accuracy · Vals AI · TaxEval ↗
MedCode ↗59.06 %
Gemini 3.1 Pro Preview (02/26) · MedCode · Accuracy ↗
MedScribe ↗76.11 %
Gemini 3.1 Pro Preview (02/26) · MedScribe · Accuracy ↗
MedQA (USMLE) ↗96.37 %
Gemini 3.1 Pro Preview (02/26) · MedQA · Accuracy ↗
ProofBench ↗26 %
Gemini 3.1 Pro Preview (02/26) · ProofBench · Accuracy ↗
AIME 2024–2025 · Vals ↗98.13 %
Gemini 3.1 Pro Preview (02/26) · AIME · Accuracy ↗
Gemini 3.1 Pro Preview (02/26) (Mini-SWE-agent) · Terminal-Bench-Science · Score ↗
Gemini 3.1 Pro Preview (02/26) · BioMysteryBench · Accuracy ↗
MysteryMechanism ↗20.72 %
Gemini 3.1 Pro Preview (02/26) · MysteryMechanism · Accuracy ↗
GPQA Diamond ↗95.45 %
Gemini 3.1 Pro Preview (02/26) · GPQA Diamond · Accuracy ↗
MMLU-Pro ↗90.99 %
Gemini 3.1 Pro Preview (02/26) · MMLU Pro · Accuracy ↗
MMMU-Pro ↗88.21 %
Gemini 3.1 Pro Preview (02/26) · MMMU Pro · Accuracy ↗
Gemini 3.1 Pro Preview (02/26) · SAGE · Accuracy ↗
Vibe Code Bench ↗6.69 %
Gemini 3.1 Pro Preview (02/26) · Vibe Code Bench · Vibe Code Bench 1–100 ↗
1 more reported result
32.03 % · Gemini 3.1 Pro Preview (02/26) (OpenHands) · Vibe Code Bench · original task set · Vals AI · Vibe Code Bench ↗
Gemini 3.1 Pro Preview (02/26) · ProgramBench · Accuracy ↗
Terminal-Bench 4.0 ↗2.52 %
Gemini 3.1 Pro Preview (02/26) (Mini-SWE-agent) · Terminal-Bench 4.0 · Score ↗
IOI v1 & IOI v2 ↗51.83 %
Gemini 3.1 Pro Preview (02/26) · IOI · Accuracy ↗
Code Migration v1.1 ↗17.31 %
Gemini 3.1 Pro Preview (02/26) · Code Migration v1 · Accuracy ↗
SWE-bench Verified ↗78.8 %
Gemini 3.1 Pro Preview (02/26) (Mini-SWE-agent) · SWE-bench Verified · Score ↗
LiveCodeBench ↗88.48 %
Gemini 3.1 Pro Preview (02/26) · LiveCodeBench · Accuracy ↗
SRE-Bench ↗0 %
Gemini 3.1 Pro Preview (02/26) · SRE Bench · Accuracy ↗
Public Benefits Bench v1.1 ↗53.79 %
Gemini 3.1 Pro Preview (02/26) · Public Benefits Bench · Accuracy ↗
1 more reported result
50.88 % · Gemini 3.1 Pro Preview (02/26) · Public Benefits Bench v1 · Accuracy · Vals AI · Public Benefits Bench v1 ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
Gemini 3 Flash (12/25) →Google
21 of 32 shared benchmark comparisons
Gemini 3.6 Flash →Google
21 of 32 shared benchmark comparisons
Gemini 3.5 Flash →Google
20 of 32 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- AIME — evaluation methodology and results ↗
- BioMysteryBench — evaluation methodology and results ↗
- CaseLaw — evaluation methodology and results ↗
- Code Migration v1 — evaluation methodology and results ↗
- CorpFin — evaluation methodology and results ↗
- Excel Modeling Benchmark v1 — evaluation methodology and results ↗
- Finance Agent v2 — evaluation methodology and results ↗
- GPQA Diamond — evaluation methodology and results ↗
- Harvey's Legal Agent Benchmark v1 — evaluation methodology and results ↗
- IOI — evaluation methodology and results ↗
- LiveCodeBench — evaluation methodology and results ↗
- LegalBench — evaluation methodology and results ↗
- Legal Research Bench — evaluation methodology and results ↗
- MedCode — evaluation methodology and results ↗
- MedQA — evaluation methodology and results ↗
- MedScribe — evaluation methodology and results ↗
- MMLU Pro — evaluation methodology and results ↗
- MMMU Pro — evaluation methodology and results ↗
- MortgageTax — evaluation methodology and results ↗
- MysteryMechanism — evaluation methodology and results ↗
- ProgramBench — evaluation methodology and results ↗
- ProofBench — evaluation methodology and results ↗
- Public Benefits Bench — evaluation methodology and results ↗
- Public Benefits Bench v1 — evaluation methodology and results ↗
- SAGE — evaluation methodology and results ↗
- Vals Science Index — evaluation methodology and results ↗
- SRE Bench — evaluation methodology and results ↗
- SWE-bench Verified — evaluation methodology and results ↗
- Tax Agent Bench — evaluation methodology and results ↗
- TaxEval — evaluation methodology and results ↗
- Terminal-Bench 4.0 — evaluation methodology and results ↗
- Terminal-Bench-Science — evaluation methodology and results ↗
- Vals Multimodal Index — evaluation methodology and results ↗
- Vibe Code Bench 1-100 — evaluation methodology and results ↗
- Vibe Code Bench — evaluation methodology and results ↗
Names used in source tables
Gemini 3.1 Pro Preview (02/26) · Gemini 3.1 Pro Preview (02/26) (Mini-SWE-agent) · Gemini 3.1 Pro Preview (02/26) (OpenHands)