OpenAI · model
GPT-5.4 Pro benchmark scores
As of 2026-11-22, GPT-5.4 Pro (OpenAI) has measured scores on 9 benchmarks and estimated scores on 76 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 93.4% | estimated ± 2.7 pp, high confidence |
| AA-HLE | 41.7% | estimated ± 4.6 pp, high confidence |
| AIME26 | 97.9% | estimated ± 1.4 pp, medium confidence |
| Artificial Analysis Intelligence Index | 33.2% | estimated ± 5.6 pp, medium confidence |
| CritPt | 30.0% | measured |
| GDPval-AA | 51.8% | estimated ± 7.8 pp, medium confidence |
| GPQA Diamond | 93.1% | estimated ± 1.4 pp, high confidence |
| HLE | 58.7% | measured |
| AA-Omniscience Accuracy | 53.2% | estimated ± 7.4 pp, medium confidence |
| BrowseComp | 89.3% | measured |
| HLE w/ tools | 61.5% | estimated ± 5.6 pp, medium confidence |
| VITA-Bench | 47.0% | estimated ± 6.7 pp, low confidence |
| τ²-bench results | 93.6% | estimated ± 11.8 pp, low confidence |
| APEX-Agents-AA | 39.7% | estimated ± 4.4 pp, high confidence |
| BioMysteryBench (human-difficult) | 49.7% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 91.6% | estimated ± 6.8 pp, low confidence |
| CyberGym | 86.7% | estimated ± 7.3 pp, medium confidence |
| JobBench | 55.3% | estimated ± 10.7 pp, low confidence |
| SkillsBench | 62.2% | estimated ± 4.1 pp, medium confidence |
| WideResearch | 82.0% | estimated ± 3.5 pp, high confidence |
| AA Agentic Index | 47.6% | estimated ± 8.2 pp, medium confidence |
| MMLU-Pro | 85.7% | estimated ± 2.5 pp, low confidence |
| FailureSensorIQ | 67.9% | estimated ± 2.1 pp, high confidence |
| GPQA | 94.3% | estimated ± 1.6 pp, high confidence |
| Vals GPQA Diamond | 93.2% | estimated ± 2.3 pp, high confidence |
| Vals MMLU-Pro | 88.4% | estimated ± 1.8 pp, high confidence |
| Chess Puzzles | 59.0% | measured |
| Gert Labs | 63.7% | estimated ± 11.1 pp, low confidence |
| AA Harvey LAB v1.0 | 92.8% | estimated ± 9.8 pp, low confidence |
| AA Tau3 Banking | 43.2% | estimated ± 3.6 pp, medium confidence |
| ARC-AGI-1 | 94.5% | measured |
| ARC-AGI-2 | 83.3% | measured |
| EBR-bench | 48.0% | estimated ± 12.7 pp, low confidence |
| FrontierMath v2 (Tier 4) | 58.5% | measured |
| FrontierMath v2 (Tiers 1-3) | 82.5% | measured |
| Furniture Assembly | 55.6% | estimated ± 14.1 pp, low confidence |
| MaxxBench | 64.9% | estimated ± 10.3 pp, low confidence |
| MazeBench | 7.2% | estimated ± 6.4 pp, low confidence |
| Mystery Game Puzzles | 52.9% | estimated ± 10.3 pp, low confidence |
| NYT Connections | 91.9% | estimated ± 12.6 pp, low confidence |
| OSWorld-Verified | 83.9% | estimated ± 5.7 pp, low confidence |
| CWE-Bench-AA | 54.9% | estimated ± 6.4 pp, medium confidence |
| AA Cyber Index | 42.1% | estimated ± 13.5 pp, low confidence |
| DeepsecBench-AA | 23.6% | estimated ± 9.0 pp, low confidence |
| HealthBench Professional (length-adjusted, Artificial Analysis) | 57.9% | estimated ± 4.4 pp, low confidence |
| HealthBench Professional (raw, Artificial Analysis) | 69.9% | estimated ± 0.8 pp, medium confidence |
| AA ITBench | 49.8% | estimated ± 4.1 pp, high confidence |
| DRACO | 85.2% | estimated ± 2.9 pp, medium confidence |
| FrontierPhysics | 14.0% | estimated ± 6.1 pp, medium confidence |
| HLE w/o tools | 42.7% | measured |
| SpireClimb | 11.2% | estimated ± 10.3 pp, low confidence |
| StealthMark | 59.0% | estimated ± 9.7 pp, low confidence |
| Toolathlon-Verified | 74.1% | estimated ± 7.4 pp, medium confidence |
| ARC-AGI-3 | 4.7% | estimated ± 7.2 pp, medium confidence |
| SkillsBench official leaderboard (BenchFlow) | 59.2% | estimated ± 6.2 pp, medium confidence |
| FACTS Multimodal | 50.5% | estimated ± 5.6 pp, medium confidence |
| ExploitBench | 78.5% | estimated ± 13.4 pp, low confidence |
| ExploitGym | 25.8% | estimated ± 9.2 pp, low confidence |
| Claw-Eval | 80.0% | estimated ± 6.4 pp, medium confidence |
| HMMT Feb 2026 | 100.0% | estimated ± 1.5 pp, low confidence |
| HMMT Nov 2025 | 95.9% | estimated ± 1.5 pp, low confidence |
| MCP Atlas | 79.5% | estimated ± 10.3 pp, low confidence |
| MMAnswerBench | 94.8% | estimated ± 2.4 pp, low confidence |
| MMLU-Redux | 97.0% | estimated ± 2.1 pp, medium confidence |
| SuperGPQA | 97.4% | estimated ± 7.2 pp, low confidence |
| Toolathlon | 59.9% | estimated ± 4.2 pp, high confidence |
| HealthBench Hard | 33.9% | estimated ± 8.6 pp, low confidence |
| MedXpertQA (Text) | 60.2% | estimated ± 8.9 pp, low confidence |
| ResearchClawBench | 19.9% | estimated ± 2.3 pp, low confidence |
| DAYJOB | 9.4% | estimated ± 6.3 pp, medium confidence |
| DAYJOB Healthcare | 6.8% | estimated ± 6.1 pp, low confidence |
| BioMysteryBench (human-solvable) | 89.5% | estimated ± 0.6 pp, medium confidence |
| HealthBench (raw) | 48.4% | estimated ± 7.6 pp, low confidence |
| HealthBench (length-adjusted) | 56.9% | estimated ± 2.5 pp, low confidence |
| HealthBench Professional | 52.0% | estimated ± 4.2 pp, medium confidence |
| HLE-Verified | 52.2% | estimated ± 1.3 pp, medium confidence |
| LABBench2 | 79.8% | estimated ± 1.7 pp, low confidence |
| BlenderWorld (text-only) | 85.7% | estimated ± 4.0 pp, medium confidence |
| Cookie-Bench | 88.6% | estimated ± 2.5 pp, medium confidence |
| Agents' Last Exam | 39.5% | estimated ± 11.6 pp, low confidence |
| SEC-Bench Pro | 69.3% | estimated ± 12.3 pp, low confidence |
| BioBasicsBench | 34.3% | estimated ± 5.1 pp, low confidence |
| LifeVal | 79.1% | estimated ± 1.6 pp, medium confidence |
| ExploitBench Internal Port | 6.7% | estimated ± 1.4 pp, medium confidence |
| PinchBench | 90.7% | estimated ± 3.6 pp, low confidence |