Calibration
JobBench → WideResearch
WideResearch is estimated from JobBench with a linear curve fitted on 5 models measured on both: y = 0.2183·x + 0.7035, R² = 0.98, cross-validated error 1.2 pp. It is used for 25 estimates.
| Estimated model | JobBench | WideResearch | Source |
|---|---|---|---|
| Claude 4.1 Opus | 21.9% | 75.1% | estimated ± 1.2 pp, medium confidence |
| Claude 4 Sonnet | 18.4% | 74.4% | estimated ± 1.2 pp, medium confidence |
| Claude Haiku 4.5 | 16.0% | 73.8% | estimated ± 1.2 pp, medium confidence |
| Claude Opus 4.6 | 36.7% | 78.4% | estimated ± 1.2 pp, medium confidence |
| Claude Opus 4.7 (Adaptive) | 45.9% | 80.4% | estimated ± 1.2 pp, medium confidence |
| Claude Sonnet 4.5 | 27.7% | 76.4% | estimated ± 1.2 pp, medium confidence |
| Claude Sonnet 4.6 | 36.9% | 78.4% | estimated ± 1.2 pp, medium confidence |
| Gemini 3 Flash | 11.4% | 72.8% | estimated ± 1.2 pp, medium confidence |
| Gemini 3 Pro | 11.4% | 72.8% | estimated ± 1.2 pp, medium confidence |
| GPT-5.1-Codex | 26.2% | 76.1% | estimated ± 1.2 pp, medium confidence |
| GPT-5.2 | 34.3% | 77.8% | estimated ± 1.2 pp, medium confidence |
| GPT-5.2-Codex | 26.0% | 76.0% | estimated ± 1.2 pp, medium confidence |
| GPT-5.3 Codex | 33.7% | 77.7% | estimated ± 1.2 pp, medium confidence |
| GPT-5.4 | 38.9% | 78.8% | estimated ± 1.2 pp, medium confidence |
| GPT-5.5 | 42.7% | 79.7% | estimated ± 1.2 pp, medium confidence |
| GPT-5 (high) | 8.5% | 72.2% | estimated ± 1.2 pp, low confidence |
| Kimi K3 | 52.9% | 81.9% | estimated ± 1.2 pp, medium confidence |
| MiMo-V2.6-Flash | 61.2% | 83.7% | estimated ± 1.2 pp, medium confidence |
| MiMo-V2.6-Pro | 62.0% | 83.9% | estimated ± 1.2 pp, low confidence |
| Muse Spark 1.1 | 54.7% | 82.3% | estimated ± 1.2 pp, medium confidence |
| Muse Spark 1.3 | 64.9% | 84.5% | estimated ± 1.2 pp, low confidence |
| Qwen3.5 Plus | 18.5% | 74.4% | estimated ± 1.2 pp, medium confidence |
| Qwen3.8-27B | 33.4% | 77.6% | estimated ± 1.2 pp, medium confidence |
| Qwen3.8-Flash-Next | 55.7% | 82.5% | estimated ± 1.2 pp, medium confidence |
| Step 5 Preview | 59.0% | 83.2% | estimated ± 1.2 pp, medium confidence |