Unverified50% confidenceFactExact time
Whisper于2022年9月开源发布，后续推出Large-v3版本改善了语音识别准确率
1
Sources
50%
Confidence
Long-term
Relevance
8/6/2026
First Seen
Sources
Related Claims
UnverifiedIn terms of accuracy, Whisper large-v3 still outperforms SpeechAnalyzer in edge cases such as complex accents, specialized terminology, and noisy environments.70% similarUnverifiedThe new version of Whisper has not been open-sourced and is reportedly significantly improved in recognition accuracy and processing speed.70% similarUnverifiedWhisper is used for speech recognition, and DALL-E 3 and Sora are used for image/video generation70% similarVerifiedWhisper is an automatic speech recognition model open-sourced by OpenAI in 2022, built on a Transformer architecture and trained on 680,000 hours of multilingual audio data.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/693907API
curl https://kongchang.com/api/v1/knowledge/claims/693907MCP
get_claim(id=693907)