Expired50% confidenceBenchmarkExact time
Soul's standard mode scored 88.8%, which jumped to 91.9% with Ultra enabled, a 2.7 percentage point increase over Max mode.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/8/2026
First Seen
Valid until: 10/6/2026(expired)
Sources
GPT-5.6编程基准91.9%：三档模型选择与Ultra模式避坑指南
bilibili费曼学AI6/28/2026
Related Claims
UnverifiedThe flagship Soul achieved a score of 91.9% on the Terminal Bench 2.1 benchmark.70% similarUnverifiedOn Health Bench Professional, Soul achieves a score of 60.5, which is 8.7 points higher than GPT-5.5.68% similarUnverifiedOn the Agent's Last Exam benchmark, GPT-5.6 Soul (extra high tier) scored nearly 54%, while Fable (max tier) scored 45%.61% similarUnverifiedOn Artificial Analysis's Omniscience hallucination metric, GPT-5.6 Sol Max reached 89%, higher than Claude Fable's 55% and GLM 5.2's 28%.61% similarUnverifiedOn front-end and visual tasks, Soul can only score six to seven points, while Fable 5 can score nine to ten points.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/230625API
curl https://kongchang.com/api/v1/knowledge/claims/230625MCP
get_claim(id=230625)