Other · model
Mercury 3 benchmark scores
As of 2026-10-29, Mercury 3 (Other) has measured scores on 6 benchmarks and estimated scores on 41 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 88.6% | estimated ± 4.4 pp, high confidence |
| AA-HLE | 33.1% | estimated ± 3.8 pp, high confidence |
| AA-SciCode | 50.5% | estimated ± 3.5 pp, medium confidence |
| CritPt | 9.9% | estimated ± 4.6 pp, high confidence |
| GDPval-AA | 24.1% | estimated ± 12.2 pp, low confidence |
| GPQA Diamond | 89.2% | measured |
| HLE | 35.9% | estimated ± 7.3 pp, medium confidence |
| IFBench | 76.8% | measured |
| AA-LCR | 73.5% | estimated ± 3.4 pp, high confidence |
| BrowseComp | 76.1% | estimated ± 6.7 pp, low confidence |
| IFEval | 94.1% | estimated ± 3.6 pp, high confidence |
| AA Coding Index | 68.9% | estimated ± 2.5 pp, high confidence |
| APEX-Agents-AA | 25.2% | estimated ± 11.5 pp, low confidence |
| DeepSearchQA | 71.3% | measured |
| SWE-bench Verified | 80.6% | estimated ± 4.6 pp, high confidence |
| SWE-bench Pro | 61.5% | estimated ± 7.5 pp, low confidence |
| τ³-bench results | 30.2% | measured |
| AA Agentic Index | 4.9% | estimated ± 5.9 pp, medium confidence |
| MMLU-Pro | 87.1% | estimated ± 5.5 pp, medium confidence |
| AA-IFBench | 78.8% | estimated ± 7.4 pp, medium confidence |
| Artificial Analysis Intelligence Index | 27.3% | estimated ± 10.6 pp, low confidence |
| GPQA | 89.1% | estimated ± 1.2 pp, high confidence |
| AA-Omniscience Accuracy | 32.5% | estimated ± 8.7 pp, medium confidence |
| MMMLU | 89.5% | estimated ± 1.1 pp, medium confidence |
| Gert Labs | 58.9% | estimated ± 11.2 pp, low confidence |
| ApprenticeBench | 6.7% | estimated ± 10.1 pp, low confidence |
| ARC-AGI-1 | 87.8% | estimated ± 2.7 pp, high confidence |
| ARC-AGI-2 | 52.7% | estimated ± 10.1 pp, low confidence |
| CursorBench 3.2 | 53.4% | estimated ± 6.4 pp, low confidence |
| OSWorld-Verified | 68.3% | estimated ± 7.1 pp, low confidence |
| Vals GPQA Diamond | 88.5% | estimated ± 2.0 pp, high confidence |
| Vals MMLU-Pro | 86.2% | estimated ± 1.6 pp, high confidence |
| Bug Hunt Bench | 15.6% | estimated ± 9.4 pp, low confidence |
| DeepSWE | 46.3% | measured |
| HLE w/o tools | 32.8% | estimated ± 4.3 pp, high confidence |
| PostTrainBench v1.1 | 18.9% | estimated ± 4.5 pp, medium confidence |
| ARC-AGI-3 | 0.0% | estimated ± 3.6 pp, high confidence |
| MMLU-Redux | 92.5% | estimated ± 2.8 pp, high confidence |
| NL2Repo | 50.6% | estimated ± 6.2 pp, medium confidence |
| SuperGPQA | 75.1% | estimated ± 13.9 pp, low confidence |
| MedXpertQA (Text) | 51.6% | estimated ± 4.5 pp, medium confidence |
| MMLU-Pro (Arcee) | 86.6% | estimated ± 3.4 pp, low confidence |
| ResearchClawBench | 15.3% | estimated ± 2.9 pp, low confidence |
| HealthBench Professional | 48.7% | estimated ± 4.8 pp, low confidence |
| MRCRv2 | 75.5% | measured |
| PinchBench | 91.5% | estimated ± 5.5 pp, low confidence |
| SWE-Atlas Codebase QnA | 39.8% | estimated ± 5.4 pp, low confidence |