Head-to-Head: Clean Evals vs. Legacy Benchmarks
| Capability Domain | Clean Evals (Tiers S & A) | Legacy Benchmarks (Tiers C & D) | Public Lift (Boost) |
|---|---|---|---|
| Agentic & Enterprise | 27.8p | 41.0p | +13.2p |
| Coding & SWE | 47.3p | 47.6p | +0.4p |
| Math & Scientific Reasoning | 17.6p | 25.0p | +7.4p |
Audited Benchmark Performance of GLM-5.3-Flash
Grouped by benchmark integrity so private capability signal stands apart from aging public tests.
Top-Tier Validated Tests (Tiers S & A) (15)
Private, leak-resistant tests that build real capability signal.
| Benchmark | Benchmark Integrity Status | Domain | Rank | Percentile |
|---|---|---|---|---|
| ITBench-AA | Private / Fresh | Coding & SWE | #3/12 | 83.1p |
| AutomationBench-AA | Private / Fresh | Agentic & Enterprise | #28/50 | 44.9p |
| AA-Briefcase v1.1 | Private / Fresh | Agentic & Enterprise | #32/50 | 36.7p |
| Harvey LAB-AA v1.1 | Private / Fresh | Agentic & Enterprise | #18/21 | 31.6p |
| MKA (Multimodal Knowledge Agent) | Private / Fresh | Agentic & Enterprise | #19/23 | 31.2p |
| FrontierCode-AA 2.0 | Private / Fresh | Coding & SWE | #35/46 | 25.6p |
| EnterpriseOps-Gym-AA | Private / Fresh | Agentic & Enterprise | #11/13 | 16.7p |
| FrontierMath | Private / Fresh | Math & Scientific Reasoning | #26/27 | 13.2p |
| MLCR-AA | Private / Fresh | Math & Scientific Reasoning | #43/46 | 8.7p |
| CyberBench v1.1 | Private / Fresh | Coding & SWE | #16/25 | 41.1p |
| Code Migration v1.1 | Private / Fresh | Coding & SWE | #15/24 | 39.4p |
| CritPt v2 | Private / Fresh | Math & Scientific Reasoning | #28/40 | 30.8p |
| LiveBench | Private / Fresh | General Knowledge & QA | #23/28 | 30.6p |
| AA-Omniscience v1.1 | Private / Fresh | General Knowledge & QA | #35/47 | 27.3p |
| AA-AnalystAgent | Private / Fresh | Agentic & Enterprise | #24/24 | 5.4p |
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 | #37/49 | 25.0p |
| VIALS-AA | Semi-Exposed | Math & Scientific Reasoning | #18/23 | 22.7p |
Legacy & Vulnerable Benchmarks (Tiers C & D) (5)
Public or aging benchmarks prone to overfitting and training-data leaks.
| Benchmark | Benchmark Integrity Status | Domain | Rank | Percentile |
|---|---|---|---|---|
| GDPval-AA v2.1 | Public / Aging | Agentic & Enterprise | #18/43 | 59.5p |
| SWE Atlas - Codebase QnA | Public / Aging | Coding & SWE | #10/15 | 50.4p |
| Terminal-Bench 4.0 | Public / Aging | Coding & SWE | #27/50 | 44.9p |
| GPQA Diamond | Public / Aging | Math & Scientific Reasoning | #13/17 | 25.0p |
| GDP.pdf | Public / Aging | Agentic & Enterprise | #39/50 | 22.4p |