§ evals · storyline

DeepSeek V4 Pro 0813 vs Claude Fable 5 on DeepSWE: Cost, Coding, and Routing

Benchmark comparison shows DeepSeek V4 Pro outperforms Claude Fable 5 on pass@4; Fable leads pass@1 but costs 90x more.

Aug 17 · 02:00:00 · primary fetch1 sourceupdated Aug 17 · 02:00:00

We ran 904 DeepSWE rollouts on DeepSeek V4 Pro 0813 and Claude Fable 5. Fable leads pass@1 at 90x the cost; Pro wins pass@4, and a Pro-first cascade hits 82.7%.

read full article on together.ai ↗

§ sources1 publication · timeline below

together.aiDeepSeek V4 Pro 0813 vs Claude Fable 5 on DeepSWE: Cost, Coding, and Routingprimary02:00:00