# Canary 180M Flash Benchmark Scores & Performance

> Canary 180M Flash has Fermion speech evaluation results. Voice evidence is display-only and does not produce a text-model score.

Canonical page: https://benchlm.ai/models/canary-180m-flash

Last updated: September 30, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | NVIDIA |
| Source Type | Open Weight |
| Reasoning Type | Non-Reasoning |
| Context Window | N/A |
| Official model card | [NVIDIA model documentation](https://huggingface.co/nvidia/canary-180m-flash) |
| Overall Score | Not scored (voice evidence only) |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Canary 180M Flash
- Variant: base
- Benchmarks covered: 0 of 486
- Coverage note: Speech and voice results appear below; no weighted text-model benchmark rows are stored.

## Speech and voice evidence

[Phonon-2 launch evaluation](/voice-benchmarks/phonon-2-launch-evaluation)

Fermion ran Phonon-2, Phonon-1, and Parakeet Redux with the Open ASR Leaderboard code on full test sets. The other accuracy rows are leaderboard figures quoted by Fermion, without independent re-verification here. Some model cards report different runs; those values are not mixed into this comparison. Voxtral's download size is an estimate at 16 bits. Throughput includes log-mel processing and excludes model load; batch-128 GPU results are separate from single-stream results. The Mac runtime comparison uses 20 dictations totaling 797 seconds at each runtime's defaults. Download size does not measure runtime memory. These results do not enter text-model or voice-agent rankings.

| Metric | Result |
| --- | --- |
| Seven-set mean WER | 5.69% |
| LibriSpeech clean WER | 1.52% |
| LibriSpeech other WER | 3.42% |
| AMI WER | 12.09% |
| Earnings-22 WER | 8.33% |
| GigaSpeech WER | 8.87% |
| SPGISpeech WER | 2.04% |
| VoxPopuli WER | 3.57% |
| Download | 737 MB |

Leaderboard row quoted by Fermion; Fermion launch table retrieved September 30, 2026. Lower WER is better; throughput excludes model load.

- [Evaluation source](https://www.fermionresearch.com/research/phonon-2/)

Open speech recognition and translation weights under CC-BY-4.0. The official model card documents English, German, French, and Spanish, 16 kHz mono audio, and inference through NVIDIA NeMo.

## Other NVIDIA Models

- [Nemotron 3 Ultra](/models/nemotron-3-ultra) - Score: 43.01
- [Nemotron 3 Nano Omni 30B A3B](/models/nemotron-3-nano-omni-30b-a3b) - Score: 30.71
- [Nemotron 3.5 Lightning 30B A3B NVFP4](/models/nemotron-3-5-lightning-30b-a3b-nvfp4) - Score: 18.94
- [Nemotron-4 15B](/models/nemotron-4-15b) - Score: 11.45
- [Nemotron 3 Nano 30B](/models/nemotron-3-nano-30b) - Score: not computed
- [Nemotron 3 Super 120B A12B](/models/nemotron-3-super-120b-a12b) - Score: not computed
- [Nemotron 3 Super 100B](/models/nemotron-3-super-100b) - Score: not computed
- [Nemotron Ultra 253B](/models/nemotron-ultra-253b) - Score: not computed
- [Cosmos3-Edge](/models/cosmos3-edge) - Score: not computed
- [Audio Flamingo 3 7B](/models/audio-flamingo-3-7b) - Score: not computed
- [Nemotron 3.5 ASR Streaming 0.6B](/models/nemotron-3-5-asr-streaming-0-6b) - Score: not computed
- [Parakeet CTC 1.1B](/models/parakeet-ctc-1-1b) - Score: not computed
- [Parakeet TDT 0.6B v3](/models/parakeet-tdt-0-6b-v3) - Score: not computed