Google · model
Gemini 3 Pro Deep Think benchmark scores
As of 2026-10-07, Gemini 3 Pro Deep Think (Google) has measured scores on 2 benchmarks and estimated scores on 21 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 89.1% | estimated ± 2.7 pp, high confidence |
| AA-HLE | 48.8% | estimated ± 4.6 pp, high confidence |
| Artificial Analysis Intelligence Index | 46.8% | estimated ± 5.1 pp, medium confidence |
| CritPt | 25.7% | measured |
| GPQA Diamond | 89.4% | estimated ± 1.5 pp, high confidence |
| HLE | 44.2% | estimated ± 6.3 pp, medium confidence |
| AA-Omniscience Accuracy | 43.4% | estimated ± 6.2 pp, medium confidence |
| BioMysteryBench (human-difficult) | 49.1% | estimated ± 6.8 pp, low confidence |
| MMLU-Pro | 88.5% | estimated ± 5.8 pp, low confidence |
| GPQA | 89.9% | estimated ± 1.6 pp, high confidence |
| ARC-AGI-1 | 83.3% | estimated ± 1.9 pp, high confidence |
| ARC-AGI-2 | 45.1% | measured |
| Vals GPQA Diamond | 88.8% | estimated ± 2.2 pp, high confidence |
| Vals MMLU-Pro | 85.7% | estimated ± 2.0 pp, high confidence |
| HLE w/o tools | 36.3% | estimated ± 6.9 pp, medium confidence |
| C-Eval | 100.0% | estimated ± 1.4 pp, low confidence |
| HealthBench Hard | 33.7% | estimated ± 9.0 pp, low confidence |
| MMLU-Pro (Arcee) | 100.0% | estimated ± 4.0 pp, low confidence |
| ARC-AGI-3 | 0.0% | estimated ± 13.8 pp, low confidence |
| BioMysteryBench (human-solvable) | 89.5% | estimated ± 0.6 pp, medium confidence |
| HealthBench Professional | 55.1% | estimated ± 4.6 pp, medium confidence |
| HLE-Verified | 35.3% | estimated ± 0.7 pp, low confidence |
| LABBench2 | 80.7% | estimated ± 2.5 pp, low confidence |