Expired50% confidenceBenchmarkExact time
On difficult math problems, both SOL and TERRA achieved perfect scores of 10, and on the adaptive fine-tuning task all three GPT-5.6 models achieved perfect scores, whereas the previous GPT-5.5 and Opus 4.7 could only score two to three points on the same task.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/11/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/9/2026(expired)
Sources
GPT-5.6三款模型深度实测：SOL/TERRA/LUNA性能与定价全解析
bilibili薛定猫AI7/9/2026
Related Claims
Partially VerifiedAll three GPT-5.6 models achieved a perfect score of 10 on the long-horizon agent fine-tuning task, while GPT-5.5 and Opus 4.7 could only score two or three points on that task.85% similarUnverifiedSam Altman stated that due to improved reasoning capabilities, the actual number of tokens consumed per task by GPT-5.5 is fewer than that of GPT-5.4.76% similarUnverifiedOpenAI officially states that GPT-5.6 Sonara consumes only one-third the tokens of GPT-5.5 when executing programming tasks of equivalent complexity.75% similarVerifiedSOL's benchmark score is nearly double that of the previous generation GPT-5.5, representing a generational leap, but still trails Fable 5 overall.74% similarUnverifiedGPT-series models have far higher token utilization than competitors, with Sol typically completing equivalent tasks using only about one-third of the token count.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/488171API
curl https://kongchang.com/api/v1/knowledge/claims/488171MCP
get_claim(id=488171)