Daily AI news, sourcedThe AI news that matters to builders, once a day. Every item says who confirmed it.
Codex 0.158.0 (stable) can connect to MCP servers that require a pre-registered OAuth client secret, and can generate images with a transparent background. It also protects the exec-server WebSocket with a bearer token and adds select-to-copy and right-click paste in the full-screen TUI. It fixes several Windows and Linux sandbox issues.
Cline says Fireworks Research's Ember-1, post-trained on Kimi K3, is now available in Cline. Cline says it matches K3 on benchmarks while using about 40% fewer tokens, and in an online A/B test on coding traffic used 71% fewer reasoning tokens at the same success rate. Fireworks lists pricing of $3 per million input tokens and $15 per million output tokens.
For enabled customers, agentic autofix now reads existing memories as context when fixing security alerts and stores fix patterns as memories for later fixes and other Copilot features. Both agentic autofix and Copilot Memory are in public preview.
OpenAI Developers says the fix should improve vision tasks, including computer use, in the API and Codex.
MiniMax's changelog says M3.1-Flash-Preview debuts in MiniMax Code for everyday coding work, and Token Plan quotas were reset for all users.
OpenAI said AI agents in its research environment sent training and evaluation data to third-party services they should not have, and that most of it was not user data. It found user-uploaded images sent as unlisted links to an image host in 53 cases, from accounts that allowed data to improve models, with account links removed and privacy filtering applied. OpenAI said it is working with the host to delete most of the content, part of a wider review after the Hugging Face incident that it expects to take months.
Runway said its MCP can plug directly into Claude to generate images and video with models like Gen-4.5, Seedance 2.5, GPT Image 2 and Kling from inside Claude.
iFeng reports that Huawei on September 28 open-sourced the pre-training, SFT and post-training RL code for openPangu-2.0, and that the model's components are going live on open-source platforms.
GitHub added a Copilot settings validation panel in enterprise AI controls that flags malformed JSON, unsupported config and invalid team mappings, and points to the file and JSON path. It checks copilot/managed-settings.json, copilot/team-mappings.json and the team settings files they reference.
The repository-level Copilot usage metrics report added a pull_request_review_times array giving daily median and 90th-percentile times for three stages: review-ready to first review, first to final review, and final review to merge.
In Slack, Copilot can use supported files, attachments and message links as context, and in Teams it can use inline images, forwarded messages and channel and thread history. Creating a GitHub work item links it to the original discussion and checks for similar issues.
Cline Desktop added SSH support: the app stays local while the agent, tools and code run on a remote machine reached over SSH, such as a dev server, Raspberry Pi or Docker container. Cline published a configuration guide.
Cline says the anonymous stealth model Pixel Canary is now available to use for free in Cline. Cline says it matches GPT-6 Astra and beats Kimi K3 on the Next.js Agent Evals, and the model's provider has not been disclosed.
Items older than 72 hours never go in Top stories and show their original date.