Home Model series Claude Fable Claude Fable · FrontierSWE · v2 · Release history Claude Fable · FrontierSWE · v2 · Release history View scores, configurations, and disclosures for FrontierSWE · v2 across 2 Claude Fable releases, ordered by release date.
Differences are calculated only for unique, active, exact values between direct predecessors when the unit, metric, and explicitly recorded configuration and conditions match.
2026-09-15 · 2 score records for this release, including 2 exact values (may include historical records)
Claude Fable 5.1 · FrontierSWE scores and configurations Metric and score Configuration and evaluation conditions Sources Mean score 56.3% Reported 56.3% · Unit % Published Unconfirmed
最高可用 effort(逐项未注明) Settings Claude Fable 5.1;官方 benchmark 表未另注的分数按该模型最高可用 effort 报告。 Unknowns 逐 benchmark 实际 effort 未完全标注。、部分 benchmark 版本、harness、运行数未在主文列出。 Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol FrontierSWE v2; 主文表格最高可用 effort,逐项档位未标。 Conditions Later comparison table reports 56.3%; System Card reports 0.57 on a 0–1 scale. Both source values are retained; underlying difference is unresolved. View this score in rankings Mean score 57% Reported 0.57 on a 0–1 scale · Unit % Published Unconfirmed
Adaptive Max Settings Claude Fable 5.1;CursorBench 主文引述明确为 max effort;后续 System Card 对 OSWorld/BenchCAD/OfficeQA 对照结果亦标 max effort。 Unknowns 不同 benchmark 的具体执行细节分别见来源章节;部分测试的采样参数或运行数未单独公布。 Tools Tool access not recorded Reasoning budget maximum Framework not recorded Protocol FrontierSWE v2; 34 ultra-long-horizon tasks; Proximal harness; five trials per task; maximum reasoning effort. Conditions Value 0.57 is the card’s rounded score on a 0–1 scale (57%); later comparison table reports 56.3%, retained separately. View this score in rankings
benchmark table > Agentic coding > FrontierSWE v2 Main text · Introducing Claude Fable 5.2 and Claude Mythos 5.2 (later comparison source) 本地留存副本, PDF p.170 Attachment · Claude Fable 5.1 and Claude Mythos 5.1 System Card 2026-11-09 · 1 score record for this release, including 1 exact value (may include historical records)
Records for the direct predecessor do not provide matching conditions for a valid difference.
Claude Fable 5.2 · FrontierSWE scores and configurations Metric and score Configuration and evaluation conditions Sources Mean score 69.9% Reported 69.9% · Unit % Published Unconfirmed
最高可用 effort(逐项未注明) Settings Claude Fable 5.2;adaptive thinking always on;发布主表按各项最高可用 effort,逐项标注的 effort 例外优先。API 默认 effort 为 high。 Unknowns 除单项明确标出的 effort 外,表格未逐 benchmark 说明实际 effort。、System Card pp.1-264 文字已完整核读;对未在对应评测章节披露的采样设置与运行数仍保留未知。 Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol FrontierSWE v2; 主文表格最高可用 effort,逐项档位未标。 Conditions 官方 benchmark table;安全 safeguards 开启。 View this score in rankings
benchmark table > Agentic coding > FrontierSWE v2 Main text · Introducing Claude Fable 5.2 and Claude Mythos 5.2