EVA-Bench - Conversation Progression: leaderboard

Metric: Conversation progression (%; LLM-judge rating on a 1-3 scale mapped to 0-1 of whether the agent moves the call toward resolution without repetition, lost context, stalling or backtracking; mean over trials; 213 scenarios in three enterprise domains, five clean-audio trials each, averaged equally over the airline, healthcare HR and IT service management domains). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 4 models tracked.

Top models

#ModelScore
1gpt-realtime-1.567.9
2Gemini 3.1 Flash Live63.6
3Ultravox-Realtime42.9
4GPT-Realtime-mini38.8

Interactive version: theaggregate.ai/benchmark?slug=eva-bench-conversation-progression · How It Works · Data refreshed daily, snapshot 2026-09-26.