In the news
DeepSeek V4 Pro 0813 vs Claude Fable 5 on DeepSWE: Cost, Coding, and Routing
Together AI · Published · 3 min read
In 30 seconds
- What happened
- DeepSeek V4 Pro costs 90x less than Claude Fable 5 per rollout while matching its accuracy on retries and exceeding it at pass@4 on DeepSWE coding tasks.
- Why it matters
- Engineers building high-volume coding agents or systems that can retry should consider this pairing; Fable alone justifies cost only for Rust and serialization work.
- Watch out
- Fable still wins first-attempt accuracy by 7 points; the models fail on different tasks, making routing effective but requiring test suite validation to escalate safely.
Listen to this summary
- claude
- deepseek
Who else ran this
The same event, reported by other publishers we follow.
The patterns behind this
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.