Unverified50% confidenceBenchmarkExact time
在Terminal Bench上，Fable 5.1的分数比Grok 4.7高出约20分，接近50%的提升
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/25/2026
First Seen
Valid until: 12/24/2026
Sources
Related Entities
Related Claims
Unverified在CursorBench 4.0基准中，Grok 4.7 Extra High得分46.3%，逼近Opus 5 Max的46.6%，成本仅6.01美元74% similarUnverified在Terminal-Bench 4.0上，Claude 5.1标准版得分55.8%，受信版达到60.9%；OS World 2.0部分完成率为77.9%72% similarUnverifiedGPT-5.4准确率为50.5%，其配对的永远说涨基准为51.6%71% similarUnverifiedFable 5在SWE Bench Pro编程测试上得分80.3分，而OpenAI的GPT 5.5仅为58.6分70% similarUnverified跑分排名不等于所有场景的实际表现，Terra跑分接近5.5但游戏任务中5.5完成度略高67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/948940API
curl https://kongchang.com/api/v1/knowledge/claims/948940MCP
get_claim(id=948940)