News
DeepSeek updates V4.1 Flash and opens its in-house search API
DeepSeek upgraded V4.1 Flash to DeepSeek-V4.1-Flash-1017 on its web app, mobile app and API, reporting higher benchmark scores with shorter runtimes, while API pricing and the deepseek-flash model name stay the same. It also opened a Search & Fetch API built on its own search engine.
On 17 October 2026 DeepSeek posted two changes in its API change log. The main one is DeepSeek-V4.1-Flash-1017, an upgrade of the V4.1 Flash model released on 10 September. The new version also replaces the model in DeepSeek's web and app products. API users keep the same calling method and the same pricing and reach the new version through the deepseek-flash model name.
DeepSeek reports higher scores than the previous V4.1 Flash on the benchmarks it lists, for example 42.3 on Terminal-Bench 4.0 (previously 31.2), 77.0 on DeepSWE v1.1 (previously 74.2) and 29.1 on ProgramBench (previously 20.3). The company says it also cut task runtime and output length, so tasks finish faster and cost less. It tuned the model for DeepSeek Harness, Tencent CodeBuddy and WorkBuddy, among other platforms, and recommends DeepSeek Harness 0.2.1-rc.1 or later.
The second change opens DeepSeek's in-house search engine to developers as a Search & Fetch API for web search and page fetching. The same engine already runs search in DeepSeek's web and API products. DeepSeek describes it as an AI-native general-purpose engine that indexes hundreds of billions of web pages, supports many languages and filters out SEO spam and low-quality AI-generated content. Usage details are in DeepSeek's Search & Fetch guide.
Sources
- DeepSeek API docsOriginal announcement
The headline and summary are ours; the facts come from the sources above. See how we verify.
Also tracked: All news · Usage reset tracker · All releases · Blog