Tencent CodeBuddy

Headlines · 2

1. 7h
   DeepSeek updates V4.1-Flash for faster tasks, opens search APIs

   DeepSeek released DeepSeek-V4.1-Flash-1017 on Oct. 17, calling it a minor update that keeps the prior architecture and 552B parameter scale while cutting task execution time and output length; it is live in the web and mobile apps and behind the existing deepseek-flash API ID at unchanged prices, with MIT-licensed weights on Hugging Face. DeepSeek's own benchmark table puts the update at 77% pass@1 on DeepSWE v1.1 versus about 74% for the September version, and a 36Kr hands-on piece cites Terminal-Bench 4.0 rising from 31.2 to 42.3 with average cost per DeepSWE task falling from roughly $0.49 to $0.20. Alongside the model, DeepSeek opened its in-house search engine as Search and Fetch APIs at 5 yuan per 1,000 successful requests and tuned the release for DeepSeek Harness, Tencent CodeBuddy and WorkBuddy; AIbase reported it topped the Code Arena WebDev leaderboard among open-source models, a claim the other reports did not corroborate.

   SiliconANGLE AI · 🔥 56
2. 2d
   DeepSeek releases DeepSeek-V4.1-Flash-1017 open-weight model

   DeepSeek released DeepSeek-V4.1-Flash-1017, an open-weight model that the testingcatalog X post says scores 77% on the DeepSWE 1.1 benchmark, ahead of GLM-5.3, Kimi K3 and models such as GPT-5.6 Sol and Opus 5. The post says the model is available on DeepSeek Chat, the API and HuggingFace, and that it is optimized for DeepSeek Harness and Tencent CodeBuddy. A DeepSeek message quoted in the post claims stronger benchmark performance, faster task completion and shorter outputs than the previous version — a "Pareto improvement in performance and cost" — but the comparisons and the "extremely cheap" description come with no independent verification or pricing figures.

   testingcatalog (X) · 🔥 1

Experience and discussion from the community

Share my Tencent CodeBuddy experienceAsk about Tencent CodeBuddy