Unverified50% confidenceBenchmarkExact time
GPT-6 Astra outperformed Fable 5.1 on the hardest benchmarks selected by Anthropic to showcase Fable 5.1 (such as Terminal Bench Science and Automation Bench), and at a lower cost.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/4/2026
First Seen
Valid until: 1/2/2027
Sources
Related Entities
Related Claims
UnverifiedGPT-6 Astra ranks first on agent and computer-operation leaderboards including Terminal Bench 4, OS World, Agents Last Exam, and Automation Bench.79% similarUnverifiedGPT-6 Astra has set new records on the Terminal-Bench 4.0 and OSWorld 2.0 benchmarks.76% similarUnverifiedOn Zapier's Automation Bench, GPT-5.6 Soul leads Fable by a narrow margin of 0.7 points at the max tier, with comparable costs.75% similarUnverifiedAccording to Anthropic's official radar chart, Opus 5.5 only falls behind GPT-6 Astra in workflow and scientific research tasks.75% similarUnverifiedIn an AI travel website landing page frontend design test, the reviewer judged GPT-6 Astra to outperform Fable 5.1.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/972622API
curl https://kongchang.com/api/v1/knowledge/claims/972622MCP
get_claim(id=972622)