DeepSeek Launches V4-Pro API with Massive Agent Upgrades, Closing Gap on Claude 5
智东西 · wechat · 2026-08-13
DeepSeek has officially released the API for V4-Pro, significantly boosting its agent capabilities to rival or surpass Claude Fable 5 in certain benchmarks. Notably, its score on the DeepSWE benchmark for complex software engineering surged from 12.8 in the preview to 62.7.
The new API (version 0813) supports JSON structured output, tool calling, and the Anthropic API. However, its concurrency limit is set at 500, and official pricing has not yet been announced, despite a previous notice about upcoming price hikes.
Some developers noted a clear discrepancy in Chain of Thought (CoT) quality between the web interface and the API, with the latter producing more rigid outputs. As benchmark scores among top models converge, inference cost is becoming the decisive factor for developers.
Related event: DeepSeek launches V4-Pro with peak pricing, API price hike sparks debate(62 posts)→
More from Models
- GLM 5.5 rumored to drop soon, coding ability expected to match frontier models — bindureddy · 2026-10-16
- Five rounds of adversarial prompting make Opus 5.5 and Argon converge on identical output — LanceCharger · 2026-10-16
- DIY benchmark ranks 18 models on subjective judgments, GLM-5.3 tops consensus match — arkuto · 2026-10-16
- Bindu Reddy: Google's Argon excels at complex doc writing at half the cost of Opus — bindureddy · 2026-10-16
- MAI-2 is 2x bigger than V4-Pro while V4.1F is 3x smaller: Microsoft's scaling ladder questioned — teortaxesTex · 2026-10-16
- Step 5 Preview hits DeepInfra day-zero: 600B MoE, 1M context, $1/M input tokens — StepFun_ai · 2026-10-16