Partially Verified80% confidenceBenchmarkExact time
In the interactive chart of OpenAI's biology foundational benchmark, running JavaScript commands via browser developer tools reveals a hidden eighth model (reportedly GPT-6.2 Astra) scoring 89% at the highest setting, while GPT-6 Astra scores only 67%.
6
Sources
80%
Confidence
Short-term (~14 days)
Relevance
11/17/2026
First Seen
Valid until: 12/1/2026
Sources
GPT-6.2 Astra Leak Hands-On: AGI Declaration and the Safety Dilemma
youtubeWorldofAI11/17/2026
Related Entities
Related Claims
UnverifiedOpenAI's system card shows that GPT-6.2 Astra missed known obstacles in about 22.28% of evaluated coding cases, compared with 11.09% for the previous-generation Astra.75% similarUnverifiedGPT-6.2 Astra shows the most significant improvement on Web development and front-end tasks, which was previously a relatively weak area for OpenAI models.74% similarUnverifiedIn experiment 7, using only the weights, GPT-6 Astra could not distinguish between the correctly answering version and the sycophantic answering version of Gemma 3 270M, but when provided with the attention patterns it correctly identified all of them.71% similarVerifiedOpenAI has launched the next-generation model GPT-6 Astra and released the first batch of test feedback to the enterprise community.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1179455API
curl https://kongchang.com/api/v1/knowledge/claims/1179455MCP
get_claim(id=1179455)