Open sourcePython · ≈1.5k stars · MIT

tau-bench

Name: tau-bench
Author: sierra-research

Code and data for Tau-Bench

sierra-research/tau-bench

Maintained by sierra-research

Tau-bench is Sierra Research's benchmark code and data for evaluating tool-using agents in conversations with simulated users.

In research (5)

GEPA-Curator: Curated Learning for Self Improving AgentsVerdict: Useful
Oct 27, 2026 · tau-bench is one of the four suites GEPA-Curator is evaluated on.
Concordat: Typed Sessions for LLM AgentsVerdict: Useful
Oct 27, 2026 · The benchmark Concordat is evaluated on. The paper shows its outcome reward misses policy violations.
AT-AT: Automatic Tool Optimization from Agent TracesVerdict: Useful
Oct 26, 2026 · tau2-bench-verified is the main evaluation benchmark.
EpiSEM-Bench: Evaluating Safety and Evidence Recovery in LLM Agents under Missing EvidenceVerdict: Useful
Oct 22, 2026 · Cited as the closest policy-compliance benchmark for tool agents; EpiSEM adds missing-evidence interventions.
Agent Behavior Maps: Locating Agents by What They Would Do, Without Running ThemVerdict: Useful
Oct 21, 2026 · Public HAL tau-bench traces are among the reference trajectories.

Go · ≈49k stars · Apache-2.0

alibaba/open-code-review

Alibaba's hybrid deterministic-plus-agent code review CLI.

Maintained by Alibaba

4 pros · 4 cons

Python · ≈39k stars · MIT

stanfordnlp/dspy

Programming, not prompting: optimize the whole pipeline.

Maintained by Stanford NLP

3 pros · 4 cons

TypeScript · ≈37k stars · Other

Featured

langfuse/langfuse

Open-source LLM engineering platform: traces, evals, prompts.

Maintained by Langfuse

4 pros · 5 cons

Relations from the Harness Hub knowledge graph, as charted in the State of Harness reports.

Maintained by · 1

Similar entries · 1

Mentioned in · 1

Verified Oct 29, 2026