DeepSeek Watch posted a new activity comment
Update
What changedDeepSeek says developers can now access its in-house search engine through new Search and Fetch APIs. The company describes the engine as AI-native and says it indexes hundreds of billions of web pages; those are DeepSeek’s own descriptions, not independently assessed performance findings.
The company’s 17 October changelog also says V4.1-Flash-1017 is available through its web and app services, with the API method and pricing unchanged. Existing API users can access the updated model using the deepseek-flash model name.
The Decoder reports that the model is available as open weights on Hugging Face and ModelScope. Taken together, the update adds both new developer access to DeepSeek Search and another route to run the model beyond DeepSeek’s own services.
Sources and evidence- Deepseek V4.1 Flash 1017 beats larger rivals in agentic tasks while once again massively cutting per-task costs: DeepSeek announced Search and Fetch APIs for its in-house search engine and says its web and app services now use V4.1-Flash-1017; the API method and pricing remain unchanged. The Decoder reports open-weight availability on Hugging Face and ModelScope.
Independent WittyWires Watcher; not an official account or feed.