What AI Update Did You Miss Today?

Fable 5 Returns, Included With Your Subscription Through July 7

Anthropic has brought the Fable 5 model (claude-fable-5) back. Through July 7, 2026, Fable is available to Claude subscribers as part of their subscription rather than only through separate usage-based billing, giving developers a limited window to use the model at no extra cost inside Claude Code.

Claude Code

Claude in Chrome Reaches General Availability

Anthropic moved Claude in Chrome out of beta and into general availability, giving all paid Claude Code plans (Pro, Max, Team, Enterprise) access to a browser extension that lets Claude see, click, type, and navigate the web directly. The extension now pairs with Claude Code's terminal workflow so developers can build in the CLI and verify the result live in the browser, including reading console errors, network requests, and DOM state. Anthropic also shipped safety hardening alongside GA, citing a drop in successful prompt-injection attacks from 23.6% to 11.2% after adding new defenses.

Claude Code

GitHub Copilot: Kimi K2.7 Code Now Generally Available

GitHub made Kimi K2.7 Code generally available in GitHub Copilot, marking the first open-weight model to join the Copilot model picker. The model, hosted by GitHub on Microsoft Azure, gives developers a lower-cost alternative for coding workflows and rolls out first to Copilot Pro, Pro+, and Max plans. Business and Enterprise administrators must explicitly enable the model through Copilot settings before their organizations can use it.

GitHub Copilot

GitHub Copilot: Browser Tools for VS Code Now Generally Available

GitHub made browser tools for GitHub Copilot agents in VS Code generally available, letting agents drive a real browser, navigate live web apps, and feed what they observe back into chat. Agents can open pages, click, type, capture screenshots, and read console errors, while built-in DevTools remain available for manual inspection. Tabs stay private by default until a developer explicitly shares one with an agent, and enterprise administrators get a dedicated toggle plus existing domain restriction controls.

GitHub Copilot

GitHub Copilot: Vision Now Generally Available

GitHub Copilot vision reached general availability, letting every subscriber attach images and PDFs directly to chat prompts so Copilot can reason about screenshots, mockups, and documents alongside code. The capability works in VS Code, github.com Copilot Chat, and the Copilot CLI, and is now on by default for Free, Pro, Pro+, Business, and Enterprise plans without any administrative action required. For Business and Enterprise users, GitHub retains attached files for roughly 24 hours to power the feature.

GitHub Copilot

Claude Code: Background Agents Now Auto-Commit, Push, and Open Draft PRs

Background agents launched from `claude agents` no longer stop to ask what to do once code work is finished in a worktree: they now automatically commit the changes, push the branch, and open a draft pull request. This removes a manual handoff step that previously required the developer to return to the session and approve next steps before the work could land anywhere reviewable. Opening the PR as a draft avoids triggering CI pipelines or notifications configured for standard pull requests, giving developers a safe, low-noise way to review autonomous agent output.

Claude Code

GitHub Copilot CLI and SDK: AI Credit Session Limits

GitHub Copilot CLI and SDK now let developers cap AI credit spending within a session, addressing a real pain point after Copilot's shift to usage-based billing. An interactive /limits command lets users view, set, or remove a cap mid-session, while the --max-ai-credits flag bounds automated, noninteractive runs so unattended agents cannot burn through a budget unchecked. The feature is in public preview for Individual, Business, and Enterprise plans.

GitHub Copilot

Claude Sonnet 5 Is Now the Default Model in Claude Code

Anthropic released Claude Sonnet 5, now the default model in Claude Code, featuring a native 1-million-token context window and introductory pricing of $2/$10 per million input/output tokens through August 31, 2026 (standard pricing is $3/$15 per Mtok). Anthropic positions Sonnet 5 as approaching Opus 4.8-level performance on agentic coding, reasoning, and tool-use benchmarks at a fraction of the cost. The model is available across all Claude plans and the Claude API under the identifier claude-sonnet-5, and ships with real-time cyber safeguards enabled by default. Early third-party testing shows mixed but substantial reactions, with some reviewers calling it one of the strongest coding models yet and others flagging inconsistent results compared to the hype.

Claude Code

Gemini Omni Flash Enters Public Preview for Developers

Google released gemini-omni-flash-preview to public preview through the Gemini API, bringing native video generation and conversational video editing to developers for the first time. The model generates 3-10 second, 720p video clips from text prompts or animates still images, and lets developers refine outputs turn-by-turn using natural language via the Interactions API. Pricing is set at $0.10 per second of generated video, matching Veo 3.1 Fast, positioning Omni Flash as a lower-cost, faster-iteration alternative for video workflows.

Gemini CLI

Nano Banana Lite (Gemini 3.1 Flash-Lite Image) Reaches General Availability

Google moved gemini-3.1-flash-lite-image ("Nano Banana Lite") to general availability, positioning it as the fastest and cheapest model in the Nano Banana image family. It generates an image in about 4 seconds at $0.034 per 1,000 images, trading multi-resolution output (it's capped at 1K) for ultra-low latency and cost. The model is aimed at high-volume, budget-sensitive image generation and editing workloads, and is designed to pair with Gemini Omni Flash: generate a still cheaply with Nano Banana Lite, then animate it with Omni Flash.

Gemini CLI

GitHub Copilot: Claude Sonnet 5 Now Generally Available

GitHub rolled out Claude Sonnet 5 as a generally available model in GitHub Copilot on June 30, 2026, giving developers a new Sonnet-class option that closes much of the performance gap with Claude Opus 4.8 while remaining significantly cheaper. The model is available across Copilot Pro, Pro+, Max, Business, and Enterprise plans, rolling out gradually across VS Code, Visual Studio, JetBrains, Xcode, Eclipse, Copilot CLI, cloud agent, the Copilot App, github.com, and GitHub Mobile. Early testers describe it as markedly more agentic than its predecessor, finishing complex multi-step tasks and self-checking its own output without being prompted. Business and Enterprise administrators must explicitly enable the model through Copilot's model policy settings.

GitHub Copilot

GitHub Copilot Agent Now Native in JetBrains AI Assistant

GitHub and JetBrains deepened their partnership on June 30, 2026, making GitHub Copilot a first-class, natively integrated agent inside JetBrains AI Assistant's agent picker, rather than requiring the separate ACP Registry setup used previously. Developers can now select GitHub Copilot directly from the agent picker in JetBrains IDEs, choose their preferred model with adjustable reasoning depth, and run multistep coding tasks that iterate autonomously. GitHub says the integration is more stable and requires no manual configuration, with Next Edit Suggestions and a Skills feature planned as upcoming enhancements.

GitHub Copilot

Warp: Tab Groups Graduate to Stable Release

Warp promoted Tab Groups out of its Preview channel and switched the feature flag to stable, making named, collapsible tab organization available to every user by default rather than opt-in testers. The rollout shipped alongside several polish fixes: per-tab colors within a group (previously a group forced one uniform color on all its members), smarter new-tab placement that respects group boundaries, and higher-contrast tab bar dividers so grouped and ungrouped tabs are easier to tell apart. The change turns one of Warp's most requested workspace-organization features, first previewed on June 9, 2026, into a permanent part of the terminal for all users.

Warp

Cursor Mobile App for iOS: Cloud Agents from Anywhere

Cursor launched a native iOS app in public beta on June 29, 2026, available to all paid subscribers on the App Store. The app lets developers launch always-on cloud agents or remotely control desktop agents from their iPhone, with voice input, slash commands, and Live Activities on the lock screen that update in real time as agents work. Cursor can alert users when an agent finishes a task, needs input, or has a pull request ready for review. A launch promotion offers 75% off Composer 2.5 runs in the mobile app through July 5, 2026.

Cursor

GitHub Copilot: Claude Opus 4.8 Fast Mode in Preview

GitHub rolled out Claude Opus 4.8 in fast mode as a preview for GitHub Copilot on June 29, 2026, delivering 2.5x faster output token speeds while maintaining the same intelligence as the standard Opus 4.8 model. The fast mode variant is priced at $10 per million input tokens and $50 per million output tokens under usage-based billing, roughly three times cheaper than the fast mode tier that shipped with Claude Opus 4.7. It is available to Copilot Pro+, Max, Business, and Enterprise users across all major IDEs and developer surfaces, with Business and Enterprise administrators required to manually enable the policy in Copilot settings.

GitHub Copilot