Open ASR · AMI-Cleaned WER ↗English short-form; normalization unversioned
7.89% WEREdge0/ARK-ASR-3B
Edge0Discovered Oct 17, 2026
Edge0/ARK-ASR-3B
3.75B parameters · 19 languages · Encoder: Whisper · Decoder: Qwen2
Measured performance
All leaderboards ↗Benchmark results
Word error rate · lower is better
Open ASR · Earnings22-Cleaned-AA-chunked WER ↗English short-form; normalization unversioned
8.07% WEREdge0/ARK-ASR-3B
Open ASR · Gigaspeech-Cleaned WER ↗English short-form; normalization unversioned
6.94% WEREdge0/ARK-ASR-3B
Open ASR · LS Clean WER ↗English short-form; normalization unversioned
1.04% WEREdge0/ARK-ASR-3B
Open ASR · LS Other WER ↗English short-form; normalization unversioned
2.37% WEREdge0/ARK-ASR-3B
Open ASR · SPGISpeech WER ↗English short-form; normalization unversioned
2.46% WEREdge0/ARK-ASR-3B
Open ASR · URGENT2024 WER ↗English short-form; normalization unversioned
13.29% WEREdge0/ARK-ASR-3B
Open ASR · Voice Arena Monsoon WER ↗English short-form; normalization unversioned
3.48% WEREdge0/ARK-ASR-3B
Open ASR · Voxpopuli-AA-Cleaned WER ↗English short-form; normalization unversioned
3.55% WEREdge0/ARK-ASR-3B
Access & cost
0 quotesPrices across providers
No listed provider price
Model specifications
- Developer
- Edge0
- Modality
- Audio
- Context
- Not reported
- Maximum output
- Not reported
- Weights
- Not reported
- First discovered
- Oct 17, 2026
More from Edge0
View all ↗Edge0Discovered 1 day ago
Edge0/ARK-ASR-0.6B
1.15B parameters · 19 languages · Encoder: Whisper · Decoder: Qwen2
1.15BAudio inputTranscription
Price not listed
Edge0Discovered 1 day ago
Edge0/Audio8-ASR-0.1B
0.324B parameters · 7 languages · Encoder: Qwen3-ASR audio encoder · Decoder: 8-layer Qwen-style causal LM
0.324BAudio inputTranscription
Price not listed