Other · model
Composer 2.5 benchmark scores
As of 2026-10-07, Composer 2.5 (Other) has measured scores on 7 benchmarks and estimated scores on 22 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-SciCode | 57.6% | estimated ± 1.7 pp, medium confidence |
| SciCode | 51.1% | estimated ± 4.5 pp, medium confidence |
| Terminal-Bench 2.1 | 75.9% | estimated ± 6.9 pp, medium confidence |
| AA Coding Index | 65.9% | estimated ± 1.0 pp, medium confidence |
| SWE-bench Verified | 81.4% | estimated ± 4.5 pp, high confidence |
| SWE-bench Pro | 52.4% | estimated ± 4.3 pp, medium confidence |
| CursorBench 3.1 | 63.2% | measured |
| CursorBench 3.2 | 56.1% | measured |
| Terminal-Bench 3.0 | 4.9% | estimated ± 6.9 pp, low confidence |
| FrontierCode 1.1 Main | 38.9% | estimated ± 1.6 pp, low confidence |
| FrontierSWE v2 | 3.5% | estimated ± 7.6 pp, low confidence |
| Vals LiveCodeBench | 78.7% | estimated ± 2.7 pp, low confidence |
| Vals SWE-bench | 79.6% | measured |
| Terminal-Bench 2.1 (Vals AI) | 58.4% | measured |
| VulcanBench v3 | 70.0% | estimated ± 0.7 pp, low confidence |
| AA Terminal-Bench 2.1 | 74.6% | estimated ± 4.4 pp, high confidence |
| AA Terminal-Bench 4.0 | 11.9% | estimated ± 14.0 pp, low confidence |
| CursorBench 4.0 | 27.7% | measured |
| DeepSWE | 56.3% | estimated ± 7.2 pp, low confidence |
| PostTrainBench v1.1 | 25.2% | estimated ± 3.3 pp, high confidence |
| ProgramBench | 68.1% | estimated ± 6.0 pp, low confidence |
| SWE Multilingual | 79.8% | measured |
| MMLU-ProX | 86.6% | estimated ± 1.7 pp, low confidence |
| NL2Repo | 45.6% | estimated ± 4.7 pp, high confidence |
| NOVA-63 | 59.2% | estimated ± 1.6 pp, low confidence |
| Terminal-Bench 2.0 | 69.3% | measured |
| React Native Evals | 80.6% | estimated ± 2.4 pp, medium confidence |
| FrontierCode 1.1 Extended | 55.1% | estimated ± 1.6 pp, low confidence |
| OpenHarmony Bench | 54.2% | estimated ± 3.2 pp, high confidence |