pragma.vision Your technology observatory

Verification register AI & Agents

Living definition

What is tau2-bench (τ²-Bench)?

LLM Tool-Use & Function Calling AI & Agents

As of
2026-06-28
Revision
2026-07-19.1
Method
v1.2.0

Definition

The term, in context

AI-assisted draft · approved dataset

tau2-bench (τ²-Bench) is a benchmark that evaluates an AI agent’s ability to hold multi-turn conversations with a simulated user while following a business’s tool-use policies correctly.

Live readiness status

Status as of the current dateline

AI-assisted assembly · derived results

As of 2026-06-28, verified readiness is 58 (claimed 80, reported 68, gap 22) — Strong evidence strength; current signals suggest Adopt with guardrails.

The readiness fact belongs to the canonical Readiness Verdict for tau2-bench (τ²-Bench).

Dataset approval

Human editorial release

pragma.vision editorial — standing authorization (operator dev@soft.house, 2026-07-17)

Dataset
2026-07-19.1
Hash
sha256:879aa251cd33f19a4cd70ef7db5753df0822d65408e7a15065212c6b421fd9d4
Approved
2026-07-19

Your opinion

Tell us anything.

What works, what doesn't, what's missing — especially about our watches, lenses, and the register itself. Anonymous is fine; leave an email if you'd like a reply.