Copilot vision is now generally available. You can attach images and PDFs directly to your chat prompts so Copilot can reason about what it sees alongside your code.

Supported file types

Where it works

Copilot vision is available across the following surfaces:

Available on all Copilot plans

Copilot vision is now available to all Copilot subscribers: Free, Pro, Pro+, Business, and Enterprise. No policy changes or admin actions are required to turn it on.

Previously, users on Copilot Business and Copilot Enterprise needed the Editor Preview Features policy enabled at the org or enterprise level. Vision is now on by default for everyone.

For users on GitHub Copilot Business and GitHub Copilot Enterprise, GitHub retains image and PDF attachments for approximately 24 hours to provide the service.

The post Copilot vision is generally available appeared first on The GitHub Blog.

Searching for pull requests using author: now shows pull requests opened by Copilot cloud agent on the user’s behalf. For example, searching with author:@me on github.com/pulls will return your own pull requests and any pull requests Copilot opened at your direction. With a single query, you can easily view and manage all the pull requests you authored, whether you authored them directly or with Copilot cloud agent.

Currently, this change applies to the github.com UI and GitHub Mobile. On July 16, we will also roll out this change to the REST API and GraphQL API.

What’s changing

• Using author:[username] or author:@me in any pull request search returns human-authored and Copilot-authored pull requests together.
• Default pull request queries and views, like the Created by me view on github.com/pulls, will automatically include Copilot-authored pull requests in the list.
• You no longer need to use multiple searches or complex queries to return all of the pull requests a user is responsible for.
• The new global pull requests dashboard attributes username with Copilot as the author of pull requests, not only Copilot. Similar changes will continue to roll out across the platform over time.

The post Copilot-authored pull requests now included in author searches appeared first on The GitHub Blog.

The GitHub Copilot app is now generally available for macOS, Windows, and Linux. It’s the desktop home for agent-driven development, built natively on GitHub.

Download the GitHub Copilot app to start your first session.

Start a session from an issue, pull request, or prompt. Run parallel sessions across repositories, each on its own branch and worktree. Review the diff, validate in the integrated terminal and browser, and open a pull request that uses your team’s existing checks and merge requirements.

Since the technical preview, we’ve also added:
* Canvases: Use bidirectional surfaces where you and the agent operate on the same plan, pull request, terminal, or browser session, so progress is visible and steerable instead of buried in chat.
* Cloud automations: Schedule recurring agent work in the cloud so it doesn’t depend on your machine being awake.
* Bring your own model and tools: Pick the model behind each session and connect external tools through MCP servers.

Want to learn more first? Visit the product page or read the docs.

Note: To access the GitHub Copilot app on a Copilot Business or Enterprise plan, your organization or enterprise admin must have the Copilot CLI enabled in policy settings.

Join the discussion within GitHub Community.

The post GitHub Copilot app generally available appeared first on The GitHub Blog.

No lies detected here from Mr. Bob Odenkirk.

Alabama man Daniel Robbins and his girlfriend Jessica Folds were found dead in the woods, their bodies lying next to each other. The truck lights were still on, the driver's side door was open, and the pair lay at the end of a drag path.

OpenAI yesterday:

We’re beginning a limited preview of the GPT‑5.6 series: Sol, our flagship model; Terra, a balanced model for everyday work; and Luna, a fast and affordable model. Terra has competitive performance to GPT‑5.5 while being 2× cheaper and Luna brings strong capability at our lowest cost.

GPT‑5.6 Sol launches with our most robust safety stack to date. We strengthened protections for higher-risk activity, sensitive cyber requests, and repeated misuse, and spent multiple weeks finding weaknesses, pressure-testing our system, and hardening it against real-world attacks.

We believe in broad access, and we plan to make GPT‑5.6 Sol, Terra, and Luna generally available in the coming weeks. As part of our ongoing engagement with the U.S. government, we previewed our plans and the models’ capabilities ahead of today’s launch. At their request, we are starting with a limited preview for a small group of trusted partners whose participation has been shared with the government, before releasing more broadly.

Stephanie Palazzolo, reporting for The Information (and posted to X) regarding an internal Q&A hosted by CEO Sam Altman:

The reason for the staggered release, Altman explained: The federal government asked it to do so. Altman said that this was the best path for widely releasing the model as soon as possible, said one of the people. In a Thursday memo, Altman told staff that the government would be “approving access customer by customer during this preview period” for GPT 5.6. He added that he hoped there would be a more general release a “couple of weeks later” if all went well. [...]

Even so, after OpenAI had shared its plans for the limited release with top government officials earlier this week, Altman still received a call from Commerce Secretary Howard Lutnick cautioning the company against launching without receiving approvals from other agencies, according to a person familiar with the call.

It is perfectly reasonable to believe that the U.S. government should have regulatory approval over frontier AI models. It’s absurd to think this should be run by an apparatchik with zero AI expertise like Commerce Secretary Howard Lutnick.

AI regulation should be thoughtful, measured, consistent, objective, and deeply informed. It should not be impulsive, impetuous, petty, uninformed, subjective, inconsistent, and transactional. The latter, however, is what we’re getting.

Link: openai.com/index/previewing-gpt-5-6-sol/