TraceAV-Bench - Speech Context: leaderboard

The Speech Context sub-task of TraceAV-Bench (2026), part of its audio-centric reasoning dimension: multiple-choice questions over long videos whose answers chain several timestamped evidence hops. The score is accuracy in percent with full audio and video input; higher is better.

Metric: Accuracy (%) on the 130 Speech Context (SC) questions, four-option multiple choice over long audio-visual videos (10 to 140 minutes), exact match of the selected option set, each model at its official inference setting, with the full audio and video input; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 3.1 Pro (Preview)96.92#99
2Gemini 3 Flash86.92#175
3Gemini 2.5 Pro83.85#276
4Gemini 2.5 Flash81.54#463
5Gemini 2.0 Flash70#723
6OmniVinci-9B65.38
7Qwen 3 Omni 30B A3B63.85#847
8Ming-Flash-Omni-2.063.85#1266
9MiniCPM-o-4.561.54#1252
10Qwen2.5-Omni-7B60#1884
11Gemma 4 E4B55.38#2184
12HumanOmni-7B52.31#3308
13video-SALMONN 247.69#2687
14VideoLLaMA2.1-7B-AV36.15#4204
15Baichuan-Omni-1.532.31#3285

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=traceav-bench-speech-context · How It Works · Data refreshed daily, snapshot 2026-11-21.