Ben Shi

Ben is on the Research team at Sierra. Previously, he was at Princeton University and Meta.

Beiträge von Ben Shi

White text "ττ-bench" on a blurry, grainy background transitioning from orange to yll.

Hyper-𝜏-bench: Evaluating agents that build agents

Today we’re open-sourcing hyper-𝜏-bench, a new long horizon agent evaluation that measures how well models can not only act as an agent, but construct one.

8. September 2026

Text "τ-knowledge" in white, with a glowing 'e', against a grainy, abstract background of shifting pink, purple, and green tones.

𝜏-knowledge: benchmarking agents on real-world knowledge

𝜏-Knowledge measures how well agents can work through messy, evolving knowledge bases to complete complex, multi-step tasks. While models are improving, they still struggle to reliably use this information in practice, leaving a large gap to real-world performance.

13. Mai 2026

The text "τ³-Bench" on a blurry, grainy background of green and brown.

𝜏³-Bench: Advancing agent benchmarking to knowledge and voice

𝜏³-Bench is here. We've expanded agent evaluation to two new frontiers: knowledge retrieval and voice.

18. März 2026

Gradient image with voice and model icons

Improving voice performance with post-training

Post-training helps our customers' voice agents achieve shorter, clearer, and more human-like conversations.

12. November 2025

Gradient mixed colors, t-bench and leaderboard icons

𝜏-Bench leaderboard: compare, explore, and understand agent performance

Introducing the 𝜏-Bench leaderboard — a community-driven platform where researchers can submit, verify, and compare results while exploring model behavior through interactive tools.

13. Oktober 2025

Ben Shi

Beiträge von Ben Shi

Hyper-𝜏-bench: Evaluating agents that build agents

𝜏-knowledge: benchmarking agents on real-world knowledge

𝜏³-Bench: Advancing agent benchmarking to knowledge and voice

Improving voice performance with post-training

𝜏-Bench leaderboard: compare, explore, and understand agent performance

Entdecken Sie, was Sierra für Sie tun kann

Beiträge von Ben Shi

Hyper-𝜏-bench: Evaluating agents that build agents

𝜏-knowledge: benchmarking agents on real-world knowledge

𝜏³-Bench: Advancing agent benchmarking to knowledge and voice

Improving voice performance with post-training

𝜏-Bench leaderboard: compare, explore, and understand agent performance