True Index: Domain Rankings
Agentic & Enterprise72.7p
#12
5 Top-Tier private testsCoding & SWE72.8p
#11
5 Top-Tier private testsMath & Scientific Reasoning91.6p
#4
3 Top-Tier private testsHead-to-Head: Clean Evals vs. Legacy Benchmarks
| Capability Domain | Clean Evals (Tiers S & A) | Legacy Benchmarks (Tiers C & D) | Public Lift (Boost) |
|---|---|---|---|
| Agentic & Enterprise | 72.7p | 63.8p | -8.9p |
| Coding & SWE | 72.8p | 76.6p | +3.7p |
| General Knowledge & QA | 94.7p | 82.8p | -11.9p |
| Math & Scientific Reasoning | 91.6p | 86.2p | -5.4p |
Non-comparable pre-release tests omitted from audited scoring (model released 2026-09-03): SWE-bench Pro V2 (2026-09-22).
Clean Evals Avg (Tiers S & A)
83.0p
16 private, leak-resistant tests
Legacy Benchmarks Avg (Tiers C & D)
77.3p
8 public or aging tests
Matched Public Lift (Boost)
-0.4p
Near zero or negative indicates clean generalization
Machine-Readable Export
Audited Benchmark Performance of GPT-6 Astra
Grouped by benchmark integrity so private capability signal stands apart from aging public tests.
Top-Tier Validated Tests (Tiers S & A) (16)
Private, leak-resistant tests that build real capability signal.
| Benchmark | Benchmark Integrity Status | Domain | Rank | Percentile |
|---|---|---|---|---|
| MLCR-AA | Private / Fresh | Math & Scientific Reasoning | #2/46 | 97.6p |
| FrontierMath | Private / Fresh | Math & Scientific Reasoning | #4/27 | 93.1p |
| FrontierCode-AA 2.0 | Private / Fresh | Coding & SWE | #6/46 | 91.0p |
| Code2Video-AA | Private / Fresh | Coding & SWE | #3/23 | 81.2p |
| MKA (Multimodal Knowledge Agent) | Private / Fresh | Agentic & Enterprise | #8/23 | 79.2p |
| Harvey LAB-AA v1.1 | Private / Fresh | Agentic & Enterprise | #8/21 | 78.8p |
| AutomationBench-AA | Private / Fresh | Agentic & Enterprise | #15/50 | 71.4p |
| AA-Briefcase v1.1 | Private / Fresh | Agentic & Enterprise | #17/50 | 67.3p |
| ARC-AGI-2 | Private / Fresh | General Knowledge & QA | #2/22 | 100.0p |
| LiveBench | Private / Fresh | General Knowledge & QA | #3/28 | 100.0p |
| Code Migration v1.1 | Private / Fresh | Coding & SWE | #4/24 | 84.5p |
| AA-Omniscience v1.1 | Private / Fresh | General Knowledge & QA | #9/47 | 84.1p |
| CritPt v2 | Private / Fresh | Math & Scientific Reasoning | #3/40 | 84.0p |
| AA-3D | Private / Fresh | Coding & SWE | #6/22 | 83.3p |
| AA-AnalystAgent | Private / Fresh | Agentic & Enterprise | #9/24 | 67.0p |
| CyberBench v1.1 | Private / Fresh | Coding & SWE | #23/25 | 24.1p |
Mid-Tier Evaluations (Tier B) (2)
Partially exposed or transitional evaluations with moderate leakage risk.
| Benchmark | Benchmark Integrity Status | Domain | Rank | Percentile |
|---|---|---|---|---|
| HLE-Diamond | Semi-Exposed | General Knowledge & QA | #7/49 | 87.5p |
| VIALS-AA | Semi-Exposed | Math & Scientific Reasoning | #4/23 | 84.1p |
Legacy & Vulnerable Benchmarks (Tiers C & D) (8)
Public or aging benchmarks prone to overfitting and training-data leaks.
| Benchmark | Benchmark Integrity Status | Domain | Rank | Percentile |
|---|---|---|---|---|
| GDP.pdf | Public / Aging | Agentic & Enterprise | #3/50 | 94.9p |
| Terminal-Bench-Science 0.1 | Public / Aging | Math & Scientific Reasoning | #6/41 | 92.4p |
| BioMysteryBench | Public / Aging | Math & Scientific Reasoning | #5/23 | 86.6p |
| Terminal-Bench 4.0 | Public / Aging | Coding & SWE | #9/50 | 83.7p |
| HealthBench | Public / Aging | Math & Scientific Reasoning | #6/16 | 79.6p |
| SWE Atlas - Codebase QnA | Public / Aging | Coding & SWE | #7/15 | 69.4p |
| GDPval-AA v2.1 | Public / Aging | Agentic & Enterprise | #28/43 | 32.7p |
| MMMU-Pro | Public / Aging | General Knowledge & QA | #3/22 | 82.8p |