OmniStarPro-RNG (Offline) - Summary Fluency: leaderboard
Narrative quality of real-time narration on OmniStarPro (Yang et al., 2026) under offline evaluation with prescribed response times: a GPT-4o judge compares the concatenated narration with the ground-truth narrative for logic, fluency, conciseness, consistency and completeness. Score on a 0-10 scale; higher is better.
Metric: Summary fluency (0-10; GPT-4o judge rating of the concatenated narration against the ground-truth narrative on logicality, fluency, conciseness, semantic consistency and completeness; responses decoded at prescribed timestamps). Source: arxiv.org. Saturation forecast: Estimated already saturated. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 5.45 | #741 |
| 2 | MiniCPM-V-2.6 | 4.13 | #3095 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=omnistarpro-rng-offline-summary-fluency · How It Works · Data refreshed daily, snapshot 2026-11-12.