τ³-bench banking
A tool-calling benchmark in the τ-bench series that simulates a bank customer service scenario, requiring agents to call APIs to complete user requests while complying with business rules, testing instruction following and tool use.
Timeline (last 90 days)
On the τ³-bench banking benchmark, EvoSkill v2 raises held-out pass rate by 5.4 percentage points over the stronger baseline.
In the τ³-bench banking experiments, when Claude Haiku 5.5 drives both the solver and the evolver, skill gains reach as high as 21.5 percentage points.
SpreadsheetBench v2 focuses on spreadsheet manipulation, and τ³-bench banking belongs to the τ-bench family of tool-calling benchmarks simulating bank customer service scenarios.
All Facts (3)
On the τ³-bench banking benchmark, EvoSkill v2 raises held-out pass rate by 5.4 percentage points over the stronger baseline.
50%UnverifiedIn the τ³-bench banking experiments, when Claude Haiku 5.5 drives both the solver and the evolver, skill gains reach as high as 21.5 percentage points.
50%UnverifiedSpreadsheetBench v2 focuses on spreadsheet manipulation, and τ³-bench banking belongs to the τ-bench family of tool-calling benchmarks simulating bank customer service scenarios.
50%