IBMAudio matchupSpeech to text
Granite Speech 5.0 470M TurboCTCvs.Granite Speech 5.0 470M TurboCTC NC
Which model is better?
Granite Speech 5.0 470M TurboCTC NC is slightly ahead in independent tests, but the lead is not certain.
Pick Granite Speech 5.0 470M TurboCTC for
Pick Granite Speech 5.0 470M TurboCTC NC for
Shared tests
Independent evaluators and leaderboards only, the newest shared snapshot per evaluator. With 95 % ranges a lead counts when they do not overlap; for hit rates, when it holds in a 95 % test. Without any uncertainty figure at most “slightly ahead”.
Hugging Face Open ASR
Granite Speech 5.0 470M TurboCTC NC slightly aheadGranite Speech 5.0 470M TurboCTC NC is clearly ahead in 0 of 15 subtests and slightly in 12, Granite Speech 5.0 470M TurboCTC in 0 clearly and slightly in 2. 1 is level.
Open ASR English short-form WER: SPGISpeech
lower is better, as of October 28, 2026
Open ASR English short-form WER: Earnings22-Cleaned-AA-chunked
lower is better, as of October 28, 2026
Open ASR leaderboard English short-set WER: Earnings-22
lower is better, as of October 16, 2026
Show all 15 subtests
| Subtest | Granite Speech 5.0 470M TurboCTC | Granite Speech 5.0 470M TurboCTC NC | Verdict |
|---|---|---|---|
| Open ASR English short-form WER: SPGISpeech (lower is better) | 3.2 % | 1.8 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: Earnings22-Cleaned-AA-chunked (lower is better) | 6.8 % | 8.2 % | Granite Speech 5.0 470M TurboCTC slightly ahead |
| Open ASR leaderboard English short-set WER: Earnings-22 (lower is better) | 6.8 % | 8.1 % | Granite Speech 5.0 470M TurboCTC slightly ahead |
| Open ASR English short-form WER: URGENT24 (lower is better) | 11.8 % | 10.2 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: Voice Arena Monsoon (lower is better) | 4.8 % | 4.3 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: Private (scripted) (lower is better) | 4.1 % | 3.8 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR leaderboard English short-set WER: AMI (lower is better) | 7.7 % | 7.1 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: Average WER (11-set macro-average) (lower is better) | 6.4 % | 6.0 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: AMI-Cleaned (lower is better) | 7.7 % | 7.3 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR leaderboard English short-set WER: GigaSpeech (lower is better) | 8.7 % | 8.3 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR leaderboard English short-set WER: VoxPopuli (lower is better) | 4.9 % | 4.7 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: Voxpopuli-AA-Cleaned (lower is better) | 4.9 % | 4.7 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: Gigaspeech-Cleaned (lower is better) | 8.7 % | 8.4 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: Private (conversational) (lower is better) | 14.0 % | 13.5 % | Granite Speech 5.0 470M TurboCTC NC slightly ahead |
| Open ASR English short-form WER: LS Other (lower is better) | 2.7 % | 2.7 % | level |
Source · latest October 28, 2026
Vendor figures
Figures from vendor reports that cover both models. They do not count toward the verdict.
Show 10 vendor figures
| Figure | Source | Granite Speech 5.0 470M TurboCTC | Granite Speech 5.0 470M TurboCTC NC |
|---|---|---|---|
| WER (lower is better) | IBM Granite Speech announcement | 5.0 % | 4.9 % |
| FFASR average WER: High SNR (lower is better) | IBM Granite Speech model card, FFASR WER figure | 10.5 % | 9.1 % |
| FFASR average WER: Lab Meas. (lower is better) | IBM Granite Speech model card, FFASR WER figure | 26.2 % | 23.1 % |
| FFASR average WER: Lab Sim. (lower is better) | IBM Granite Speech model card, FFASR WER figure | 24.9 % | 22.1 % |
| FFASR average WER: Low SNR (lower is better) | IBM Granite Speech model card, FFASR WER figure | 35.7 % | 32.9 % |
| FFASR average WER: Mid SNR (lower is better) | IBM Granite Speech model card, FFASR WER figure | 19.3 % | 16.7 % |
| FFASR average WER: Mov. High (lower is better) | IBM Granite Speech model card, FFASR WER figure | 11.5 % | 10.2 % |
| FFASR average WER: Mov. Low (lower is better) | IBM Granite Speech model card, FFASR WER figure | 37.2 % | 33.4 % |
| FFASR average WER: Mov. Mid (lower is better) | IBM Granite Speech model card, FFASR WER figure | 21.5 % | 18.7 % |
| FFASR average WER: Near Field (lower is better) | IBM Granite Speech model card, FFASR WER figure | 4.8 % | 4.3 % |
All data side by side
Prices are verbatim from the provider, because audio models bill per minute, character or token.
| Granite Speech 5.0 470M TurboCTC | Granite Speech 5.0 470M TurboCTC NC | |
|---|---|---|
| Task | Speech to text | Speech to text |
| Developer | IBM | IBM |
| Status | Open weights | Open weights |
| Released | August 25, 2026 | August 25, 2026 |
| Weights | open | open |
| Runs locally | yes | yes |
| Parameters | 0.47 B | 0.47 B |
| License | Apache-2.0 | CC-BY-NC-SA-4.0 |
| Popularity (out of 100) | 52 | 50 |