VoiceCodeBench leaderboard
As of 6 Nov 2026, the top score on VoiceCodeBench is 67.7% by GPT Live Transcribe (OpenAI), ahead of Grok Voice Transcribe 2.0 at 65.3%, across 20 model configurations with a published result (Vals AI, result dated 24 Sep 2026). Speech-recognition models transcribe audio containing domain entities such as drug names or ticker symbols; scored on getting those entities right. VoiceCodeBench is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- GPT Live Transcribe holds the top VoiceCodeBench score at 67.7% (Vals AI, 24 Sep 2026).
- Grok Voice Transcribe 2.0 is second on VoiceCodeBench at 65.3% (Vals AI, 24 Sep 2026).
- VoiceCodeBench: Speech recognition of entities across eight workflow domains. Run by Vals AI (Vals AI).
- 20 model configurations have a published VoiceCodeBench result on BenchLeader (BenchLeader, 6 Nov 2026).
- VoiceCodeBench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Speech recognition of entities across eight workflow domains. Run by Vals AI.
- Published by
- Vals AI
- Category
- Multimodal
- Index weight
- Reference only
- Models
- 20
- Data as of
- 6 Nov 2026
Vals AI (vals.ai).
What the test looks like
Speech-recognition models transcribe audio containing domain entities such as drug names or ticker symbols; scored on getting those entities right.
How it is scored
Entity accuracy, run by Vals AI.
What to keep in mind
A speech benchmark: only speech models are listed, and it does not measure reasoning.
- 1GPT Live Transcribe67.7%
- 2Grok Voice Transcribe 2.065.3%
- 3Ink 262.0%
- 4GPT 4o Transcribe59.7%
- 5Nova 3 Streaming59.3%
- 6Resonant 159.0%
- 7Inworld Stt 157.3%
- 8Muse Voice Transcribe57.0%
- 9Whisper Large v3 Turbo57.0%
- 10Flux General En55.7%
- 11Chirp 355.3%
- 12Scribe v2 Realtime55.3%
- 13Whisper Large v354.3%
- 14Voxtral Mini 260253.0%
- 15Grok Voice Transcribe 1.052.0%
| # | |||||
|---|---|---|---|---|---|
| 1 |  GPT Live Transcribe | 67.7% | – | $0.019 | 2026-09-24 |
| 2 |  Grok Voice Transcribe 2.0 | 65.3% | – | $0.0037 | 2026-09-24 |
| 3 | Ink 2 | 62.0% | – | – | 2026-09-24 |
| 4 | GPT 4o Transcribe | 59.7% | – | $0.0031 | 2026-09-24 |
| 5 |  Nova 3 Streaming | 59.3% | – | $0.0054 | 2026-09-24 |
| 6 | Resonant 1 | 59.0% | – | $0.0077 | 2026-09-24 |
| 7 |  Inworld Stt 1 | 57.3% | – | $0.0027 | 2026-09-24 |
| 8 |  Muse Voice Transcribe | 57.0% | – | $0.0034 | 2026-09-24 |
| 9 | Whisper Large v3 Turbo | 57.0% | – | $0.0007 | 2026-09-24 |
| 10 | Flux General En | 55.7% | – | $0.0072 | 2026-09-24 |
| 11 |  Chirp 3 | 55.3% | – | $0.018 | 2026-09-24 |
| 12 |  Scribe v2 Realtime | 55.3% | – | $0.0073 | 2026-09-24 |
| 13 | Whisper Large v3 | 54.3% | – | $0.0021 | 2026-09-24 |
| 14 |  Voxtral Mini 2602 | 53.0% | – | $0.0034 | 2026-09-24 |
| 15 | Grok Voice Transcribe 1.0 | 52.0% | – | $0.0037 | 2026-09-24 |
| 16 | GPT 4o Mini Transcribe Streaming Response | 51.7% | – | $0.0016 | 2026-09-24 |
| 17 | Voxtral Mini Transcribe Realtime 2602 | 46.0% | – | $0.0067 | 2026-09-24 |
| 18 |  Transcribe 03 2026 | 44.7% | – | – | 2026-09-24 |
| 19 |  Universal Language Model | 37.3% | – | – | 2026-09-24 |
| 20 | Universal 3.5 Pro | 30.3% | – | $0.0085 | 2026-09-24 |
More multimodal benchmarks · All LLM benchmarks
Related benchmarks
Cite as: BenchLeader, “VoiceCodeBench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vals_voice_code_bench.
VoiceCodeBench: questions
- What does VoiceCodeBench measure?
- Speech-recognition models transcribe audio containing domain entities such as drug names or ticker symbols; scored on getting those entities right. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on VoiceCodeBench?
- As of 6 Nov 2026, the highest score on VoiceCodeBench is 67.7% by GPT Live Transcribe (result dated 24 Sep 2026), ahead of Grok Voice Transcribe 2.0 at 65.3%. Scores come from Vals AI and are re-read daily.
- Which AI model leads VoiceCodeBench?
- GPT Live Transcribe leads VoiceCodeBench with 67.7% as of 6 Nov 2026 (result dated 24 Sep 2026), ahead of Grok Voice Transcribe 2.0 at 65.3%.
- How many models have VoiceCodeBench results?
- 20 model configurations have a VoiceCodeBench result on BenchLeader, all taken from Vals AI.
- Who runs VoiceCodeBench and how often is it updated?
- VoiceCodeBench is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does VoiceCodeBench count toward the BenchLeader Index?
- No. VoiceCodeBench is shown for reference but left out of the composite index.