benchgap
Calibration

Toolathlon-Verified → WideResearch

WideResearch is estimated from Toolathlon-Verified with a Michaelis–Menten curve fitted on 6 models measured on both: y = 1.4847·x / (0.57158 + x), R² = 0.91, cross-validated error 3.9 pp. It is used for 6 estimates.

Estimated modelToolathlon-VerifiedWideResearchSource
GLM-5.3-Flash78.4%85.9%estimated ± 3.9 pp, low confidence
Claude Fable 5.177.8%85.6%estimated ± 3.9 pp, low confidence
Claude Opus 5.577.8%85.6%estimated ± 3.9 pp, low confidence
Claude Sonnet 5.577.8%85.6%estimated ± 3.9 pp, low confidence
GLM-5.373.0%83.3%estimated ± 3.9 pp, medium confidence
Laguna S 2.149.7%69.1%estimated ± 3.9 pp, medium confidence