DeepSeek

DeepSeek has quietly swapped the model behind its Flash API and put a price on its own search engine

The deepseek-flash endpoint now serves V4.1-Flash-1017 with no migration step. Free calls through DeepSeek Harness, Claude Code and Codex end on 1 November.

Microsoft has finished pre-training a 2.8-trillion-parameter model, and the efficiency headline is a fitted estimate

The model is in post-training with no checkpoint, price or API announced. Microsoft plans to license the weights to partners through Foundry.

An AI agent solved 49 of 50 unpublished physics problems, and cutting off its internet access cost it three

TheoPhysEval was built from problems working physicists had not published, precisely so no model could have seen them. The best agent missed one, at an average cost of about $2.33 a problem.

Ten AI developers changed how they handle British people's data after two years of ICO supervision, and the eleventh is under investigation

The UK regulator closed a two-year supervision programme on 8 October, and opened a six-week call for evidence on agentic AI that runs to 20 November.

The attacker who hit South Korean finance left the agent's own session logs on an open server, and its author is now closing the source

CrowdStrike found the campaign because the operator's session histories, memory files and configuration were exposed in an open directory.

Mistral's new trillion-parameter model does the security work Claude and GPT-6 refuse

ML4 is a 1 trillion-parameter open-weight model trained on 3,800 Nvidia chips in Mistral's own European data centres. The number the company leads with is 82% on a vulnerability test where, it says, Claude Opus 5.5 and GPT-6 Astra score near zero because they decline the task.