DeepSeek releases open-weight V4.1-Flash-1017 and opens its search engine to developers
The company says the update beats Claude Opus 5 and GPT 5.6-Sol on its internal coding and agent benchmarks while producing shorter outputs.
Chinese AI company DeepSeek has released DeepSeek-V4.1-Flash-1017, an updated version of its Flash model. It is available through the company's website, app and API, and the weights are published on Hugging Face and ModelScope. DeepSeek says the model was trained with reinforcement learning using self-feedback. Compared with its direct predecessor, the company says it scores higher on benchmarks, finishes tasks faster and produces shorter outputs. DeepSeek calls the update a Pareto improvement. It says the model performs well in end-to-end software development and agent tasks. In DeepSeek's internal benchmarks, the update beats larger models, including Claude Opus 5 and GPT 5.6-Sol, on coding and agent tests. These results come from the company itself and have not been independently confirmed in the supplied reporting. DeepSeek is also opening its in-house search engine to outside developers for the first time through new Search and Fetch APIs. For developers, this is a lower-cost open-weight option for agent workloads, and it adds search tools that agents can call directly. The release also keeps up competitive pressure from open-weight models built by Chinese labs.