LLM Reference
activeAudio

Big Bench Audio

Metric: Accuracy (%) (higher is better)Introduced: 2025

Speech reasoning benchmark testing whether a voice model understands and reasons over what is said, not just transcribes it. Covers 12 tasks including emotion recognition, spoken arithmetic, factual Q&A, and instruction following. Evaluated by Artificial Analysis on speech-to-speech models. Higher is better. High benchmark score alone doesn't make a model the right pick — weigh it against pricing, API availability, and release date.

Models ranked

5

tracked on this benchmark

Score band

97.6 – 29.2

best → lowest tracked

Snapshot trend

-20.13

May 24 → Jun 7 · 4 models

Leaderboard

Tracked models ranked by Accuracy (%) (higher is better).

Compare candidates
#Model variant and provenanceScore
1
StepAudio 2.5 Realtime
Version: pctHarness: Not recordedEvaluator: Not recordedObserved: May 24, 2026Confidence: Not recordedSource
97.6
2
Grok Voice Think Fast 1.0
Version: pctHarness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
97.1
3
GPT Realtime 2
Version: highHarness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
96.6
4
Amazon Nova 2 Sonic
Version: pctHarness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
87.0
5
Nemotron 3 VoiceChat
Version: pctHarness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
29.2

How to read this benchmark

This benchmark scores models where higher is better. Scores are useful for directional filtering and shortlisting — not for universal quality ranking. Prefer benchmarks closest to your workload, then validate the linked model pages for pricing, context window, and provider availability.

Trust this score when

  • There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
  • The model list covers the same version family you can actually deploy today.
  • Top candidates overlap with your required routing and feature requirements.

Be cautious when

  • There is only one benchmark snapshot or the dataset appears stale.
  • The benchmark metric direction is opposite of your decision objective.
  • The score difference between options is narrow and likely within implementation variance.

FAQ

What does the Big Bench Audio benchmark measure?

Speech reasoning benchmark testing whether a voice model understands and reasons over what is said, not just transcribes it. Covers 12 tasks including emotion recognition, spoken arithmetic, factual Q&A, and instruction following. Evaluated by Artificial Analysis on speech-to-speech models. Higher is better. On this page it lists 5 tracked model variants where higher is better.

Is a higher Big Bench Audio score always better?

For this benchmark, higher is better. A high score helps you shortlist, but confirm pricing, context window, and provider availability on each model page before committing — the top scorer is not always the right pick for your workload or budget.

How current is this Big Bench Audio data?

This benchmark was last reviewed on Jun 7, 2026. The tracked score average moved -20.13 points across the last 2 snapshots.

Related benchmarks

Last reviewed: Jun 7, 2026

Resources