DeepSeek releases official V4-Pro-0813 model
Start here
DeepSeek released DeepSeek-V4-Pro-0813, the official version that supersedes the V4-Pro preview. It keeps the same model structure but adds a DSpark speculative decoding module. DeepSeek reports large gains over the preview on agentic benchmarks: Terminal Bench 2.1 (87.9 vs 72.1), DeepSWE (62.7 vs 12.8) and NL2Repo (61.5 vs 38.5), and says the model is broadly competitive with the strongest proprietary models. The release ships without a Jinja chat template, providing an encoding folder instead. The reasoning_effort parameter now has low, high and max levels controlling how much deliberation the model does before answering; DeepSeek recommends a maximum output length of 384K tokens for the high and max levels. Weights are MIT licensed, and vLLM and SGLang launch commands are documented.
Generated from reports · updated 13 h ago
Report timeline
Follow the reports to see the event from different sides.
- Hacker News · AIPicksDeepSeek releases V4-Pro-0813 with DSpark speculative decoding and stronger agentic benchmarks
DeepSeek released DeepSeek-V4-Pro-0813, the official version superseding the V4-Pro preview, built on the same model structure with an added DSpark speculative decoding module. The company reports gains over the preview on benchmarks including Terminal Bench 2.1 (87.9 vs 72.1), DeepSWE (62.7 vs 12.8) and NL2Repo (61.5 vs 38.5), and says the model is broadly competitive with the strongest proprietary models. The release ships without a Jinja chat template, instead providing an encoding folder, and adds low, high and max levels to the reasoning_effort parameter; weights are MIT licensed and vLLM and SGLang launch commands are documented.
Heat for this event
Not enough continuous observations yet to draw a trend.