AI TrackerDeepSeek news,
as it happens.
Every DeepSeek launch, price change, limit change and retirement, summarized with what you need to do about it.
DeepSeek's Search and Fetch tools are live on Vercel's AI Gateway
Vercel says DeepSeek's Search and Fetch tools are now available on its AI Gateway. They let any tool-calling model search the web and read pages, and Vercel says that switching model providers leaves the tools unchanged. Vercel's changelog covers the details.Artificial Analysis: DeepSeek V4.1 Flash on GB300 NVL72 is up to 1.5x faster per agent in a month
Artificial Analysis says DeepSeek V4.1 Flash speeds on NVIDIA GB300 NVL72 have risen by up to 1.5x tokens per second per agent over the past month on AA-AgentPerf, its benchmark of datacenter systems running realistic agentic coding workloads.DeepSeek Harness adds Creator Mode for turning ideas into plugins
The DeepSeek Harness account announced Creator Mode on X on November 2: a mode for turning ideas into plugins you can use. The demo video shows a design workspace, a sandbox game and a 3D visualizer of an LLM as examples. In a follow-up post the account links a collection of the English and Chinese prompts behind the plugins in the video.
OpenCode says its DeepSeek V4.1 Flash configs match DeepSeek's API, after reports the model feels dumber
OpenCode's founder (thdxr) wrote that OpenCode Go gets occasional reports that DeepSeek V4.1 Flash "feels dumber" for some people. He said he ran every benchmark he could find against the different configurations OpenCode Go serves, and that the results are in the same range.OpenCode Go raises its DeepSeek V4.1 Flash quota to $100 of usage
OpenCode posted "Operation Cheepseek Phase 4": its OpenCode Go subscription now includes $100 of usage for DeepSeek V4.1 Flash. The post does not say what the quota was before or whether it applies to every plan.DeepSeek V4.1 Flash is now available in Cursor
Cursor announced on October 26 that DeepSeek V4.1 Flash is available in the editor; it can be turned on under Cursor Settings > Models. Cursor says every effort level sits on the CursorBench cost frontier, at $0.60 to $0.98 per task, with no model scoring higher at that cost. These are Cursor's own figures from its own benchmark.SGLang says it now matches DeepSeek's DeepInfer engine on DeepSeek-V4.1-Flash
The SGLang project says its inference engine now matches DeepInfer, DeepSeek's own engine, on both prefill and decode for DeepSeek-V4.1-Flash. A few hours after vLLM made a similar claim, SGLang reported up to 1.04× DeepInfer's prefill throughput on 4× GB300 and 4× B300, and up to 7.4% decode throughput on 64× GB300.
vLLM says it now matches DeepSeek's DeepInfer engine on DeepSeek-V4.1-Flash
The vLLM project says its inference engine now matches DeepInfer, DeepSeek's own engine for DeepSeek-V4.1-Flash, on DeepSeek's decode configuration across batch sizes 8 to 256. It says it does so at DEP32 (data-parallel attention with expert parallelism over 32 GPUs), against DeepInfer's EP64DP64 setup.