Expired50% confidenceBenchmarkExact time
In the creator's self-made code audit benchmark, the Full version performed close to Astra and Fable, while the base version Muse Wonder scored almost only half of the Full version, a gap of more than 30%.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/13/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/27/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedMuse Glimmer 30B scored 51.2 on the SWE benchmark, meaning it can solve approximately half of real-world software engineering problems.60% similarUnverifiedIn the author's self-designed code review benchmark, Astra ranked first, Grok 4.7 came in second, and Fable 5.1 offered only 5 suggestions but of very high quality.57% similarUnverifiedOn the Frontier Code programming benchmark, Opus 5.5 beat Astra using the default tier, with a per-task cost of only one-fifth that of its competitor.57% similarUnverifiedIn a small-scale test published by the PartsPlan developer, all 30 chart values were answered correctly.56% similarUnverifiedJane Street evaluated Astra as achieving SOTA on their internal programming benchmark and surpassing Fable, but noted only a clear improvement in trading intuition (with Fable 5.1 slightly ahead in trading intuition).56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1017387API
curl https://kongchang.com/api/v1/knowledge/claims/1017387MCP
get_claim(id=1017387)