Anthropic reports 45.3% for Claude Fable 5.2 on FrontierCode 2.0 at High effort. Anthropic says this is about level with Opus 5.5's best score at roughly half the cost per task.
Tasks: Whether model code changes would be merged by open-source maintainers, as Anthropic describes it.
Setting: High effort, as stated in the chart description. Numeric budgets not stated.
Limits: Provider-reported by Anthropic in its launch announcement; not independently reproduced here. Anthropic points to the Fable 5.2 and Mythos 5.2 system card for methods; that card was not opened, so task subset, sample size, harness, prompts, trial counts and run dates are unverified. Results under the same benchmark name can be shown together as reported scores, but unmatched settings and missing methods rule out a controlled ranking. Anthropic labels the cost comparison early testing; per-task cost figures are not given in the text.
Anthropic
