Expired50% confidenceBenchmarkExact time
The review results show that Opus 5 frequently kills the wrong process (often mistakenly killing the T3 Code instance it is running in), and Sol (GPT) has approximately a 40% probability of submitting a draft PR.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
8/22/2026
First Seen
Valid until: 9/5/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedThe median output token count per problem-solving attempt for Opus 4.7 is 60,000, while GPT 5.5 requires only 16,000.72% similarUnverifiedIn the Agents' Last Exam benchmark, GPT-5.6 Sol achieved a score of 53.6, which is 13.1 points higher than Claude Fable 5 using adaptive reasoning.72% similarUnverifiedClaude Opus 4.8在AI Coding基准测试中表现超越GPT-5.570% similarUnverifiedOpus 4.7 showed a significant regression in long-context (500K–1M token) retrieval, far underperforming DeepSeek and GPT 5.5.67% similarUnverifiedAccording to unofficial leaks, Claude Opus 5 Max topped the Terminal Bench 3.0 benchmark with a score of 43.5%, surpassing GPT-5.6-Sol's 34.6%.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/788387API
curl https://kongchang.com/api/v1/knowledge/claims/788387MCP
get_claim(id=788387)