PERIT-S2S · Logistics & delivery

Driver and dispatch calls from vehicles, warehouses and doorsteps. Half of it is hands-free.

Audio comes from dispatch radio, driver hotlines, delivery exceptions. Scored on task success, with a working logistics & delivery operator as the control.

1,100
Clips in the target held-out set
79h
Audio, never published
6
Languages in the set
1
Working logistics & delivery operator as control

Leaderboard

Illustrative figures. Public runs begin when the held-out sets close.
ModelTask success (%)Median turn latencyBarge-in recovery
Human control · sector agentHuman control
93.4 ± 2.0
641ms96.4%
Frontier realtime · lab AFrontier
59.1 ± 1.7
790ms68.4%
Frontier realtime · lab BFrontier
57.2 ± 1.3
834ms59.8%
Frontier multimodal · lab CFrontier
50.1 ± 0.7
884ms53.9%
Cascaded stack (ASR → LLM → TTS)Specialist
48.2 ± 1.0
1519ms39.2%
Open weights · 8B speechOpen weights
24.6 ± 0.6
1012ms29.3%
Task families

What the set actually asks for.

  • Turn-taking & barge-in
  • Task completion
  • Instruction adherence
  • Disambiguation & repair
  • Escalation & refusal
  • Tone under pressure
Method & limits

How a run on this set is produced.

  1. 01Callers are scripted personas played by trained speakers, not synthesised — including the ones who talk over the agent.
  2. 02Every call is scored against the rubric by two reviewers; a third adjudicates splits.
  3. 038 calls per persona per model. Reported intervals are bootstrapped at 95%.
  4. 04Human control is a licensed agent from that sector handling the same personas blind.

Limits, stated up front: 6 languages only, single-channel audio, and no cross-sector transfer. The control is one operator per sector, so the human line carries its own error bar — it is a working standard, not a ceiling.