Expired50% confidenceBenchmarkExact time
The previous Advanced Voice Mode scored only 45% on GPQA.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/20/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/18/2026(expired)
Sources
Related Claims
VerifiedOn the GPQA test, the older advanced voice mode achieved approximately 45% accuracy, while GPT Live mini and the mid-tier version reached 70% to 80%81% similarUnverifiedIn the BrowseComp test, the old Advanced Voice Mode scored less than 1%, while GPT-Live 1 Mini (free) scored approximately 30%.69% similarUnverifiedSpeech generated by end-to-end TTS models based on deep learning typically achieves a naturalness MOS score of 4.0–4.5 (out of 5), approaching the level of human recordings.61% similarUnverifiedModern voice cloning solutions (such as SV2TTS, VALL-E, and CosyVoice) have already compressed the required voice samples to a few seconds.60% similarUnverifiedGPT-4o早期语音模式曾因过度模仿用户声音（voice cloning风险）及生成过于拟人化情感回应而受到批评59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/575465API
curl https://kongchang.com/api/v1/knowledge/claims/575465MCP
get_claim(id=575465)