@ArizePhoenix: - 55.8% on Terminal-Bench 4.0 (up from 42.0% for Fable 5) - 52.6% on Terminal-Bench-Science (up from 24.7%) - 65.0% on …
Summary
ArizePhoenix shows significant performance improvements on benchmarks like Terminal-Bench 4.0 and Humanity's Last Exam, with gains from previous versions such as Fable 5.
View Cached Full Text
Cached at: 09/02/26, 01:49 AM
- 55.8% on Terminal-Bench 4.0 (up from 42.0% for Fable 5)
- 52.6% on Terminal-Bench-Science (up from 24.7%)
- 65.0% on Humanity’s Last Exam with tools.
Similar Articles
@ArizePhoenix: The result: a 50% improvement over 5.2 on their in-house Code Bench, Terminal-Bench 3.0 from 4.6 → 28.3, DeepSWE v1.1 f…
DeepSWE v1.1 shows a 50% improvement on in-house Code Bench and sets new state-of-the-art scores on Terminal-Bench 3.0 and Agents' Last Exam, with emerging cyber capabilities advancing faster than expected.
ProgramBench result for Fable 5 is in, doubling Opus 4.8 even with 4.8 fallback "99% of the runs"
ProgramBench results show Fable 5 achieving double the performance of Opus 4.8, even with fallback to 4.8 in 99% of runs.
Terminal Bench 4.0 just dropped, GLM-5.3 is at the same level as Fable 5, accounting for margin of error
Terminal Bench 4.0 has been released, comparing AI models like GLM-5.3 and Fable 5, with a focus on rapid iteration to combat benchmark saturation and raising questions about cost-effective alternatives for evaluating coding agents.
Fable 5 below even Gemini 3.1 on Livebench
A discussion on LiveBench results showing Fable 5 performing below Gemini 3.1, questioning whether the benchmark is flawed or Anthropic is optimizing for benchmarks.
@FinanceYF5: Third-party tests indicate that the returned Fable 5 is not at the same level as the June release. BridgeMind retested the July 1 version with BridgeBench: [Debugging] 86.2→25.9, [Refactoring] 73.6→38.4, [Hallucinat…
Third-party tests show that after returning, the Fable 5 model's performance dropped significantly on the BridgeBench benchmark. Debugging score fell from 86.2 to 25.9, Refactoring from 73.6 to 38.4, Hallucination from 75.9 to 61.7. It is speculated that the new safety guardrails caused many tasks to be handed off to Opus 4.8.