Unverified50% confidenceBenchmarkExact time
在深度代码审查任务上Grok 4.7和Sol表现优于Opus和Fable，但Opus从5到5.5在该bench上的提升接近翻倍
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/25/2026
First Seen
Valid until: 12/24/2026
Sources
Related Entities
Related Claims
UnverifiedGrok 4.6在UI设计测试中表现平庸，逊于Fable和5.6 Sol71% similarUnverifiedOpus 5.5在token效率上不如Opus 5：Opus 5的max档每任务用73k token，Fable用78k，Opus 5.5的max档接近12万token，而GPT-6 Astra只需27k70% similarUnverified在盲测中Sol给Opus方案打8.3分、给Fable方案打6.0分69% similarUnverifiedGrok 4.5在SWE Bench Pro任务上的输出Token比Opus 4.8少4.2倍69% similarUnverifiedGrok 4.6的智能指数比Grok 4.5高5分，比Grok 4.3高23分69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/948931API
curl https://kongchang.com/api/v1/knowledge/claims/948931MCP
get_claim(id=948931)