Release notes
Official updates from Anthropic, OpenAI, Google, xAI, Mistral, DeepSeek and Meta, newest first. Open any update for a short AI summary and the full notes.
Google · Gemini CLI·Dec 22, 2025
fallback to GEMINI_CLI_IDE_AUTH_TOKEN env var. display quota stats for unused models in /stats.
Anthropic · Claude Code·Dec 19, 2025
Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation. Added /terminal-setup support for Kitty, Alacritty, Zed, and Warp terminals.
Google · Gemini CLI·Dec 19, 2025
cherry-pick 9e6914d to release/v0.21.2-pr-15288 to patch version v0.21.2 and create version 0.21.3.
Added clickable [Image #N] links that open attached images in the default viewer. Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank.
Anthropic · Claude API·Dec 19, 2025
Read more in Model deprecations.
OpenAI · Codex app·Dec 19, 2025
Codex now supports agent skills: reusable bundles of instructions (plus optional scripts and resources) that help Codex reliably complete specific tasks. Skills are available in both the Codex CLI and IDE extensions. You can invoke a skill explicitly by typing $skill-name (for example, $skill-installer or the experimental $create-plan skill after installing it), or let Codex select a skill automatically based on your prompt. Learn more in the skills documentation... Folder-based standard (agentskills.io). Following the open agent skills specification, a skill is a folder with a required SKILL.md and optional supporting files: text my-skill/ SKILL.md # Required: instructions + metadata scripts/ # Optional: executable code references/ # Optional: documentation assets/ # Optional: templates, resources. Install skills per-user or per-repo. You can install skills for just yourself in ~/.codex/skills, or for everyone on a project by checking them into.codex/skills in the repository. Codex also ships with a few built-in system skills to get started, including $skill-creator and $skill-installer. The $create-plan skill is experimental and needs to be installed (for example: $skill-installer install the create-plan skill from the.experimental folder). Codex ships with a small curated set of skills inspired by popular workflows at OpenAI. Install them with $skill-installer, and expect more over time.
Google · Gemini API·Dec 19, 2025
Introduced a breaking change to the Interactions API in v1beta. The total_reasoning_tokens field has been renamed to total_thought_tokens to better align with the concept of "thoughts" in thinking models.
Google · Antigravity·Dec 19, 2025
Higher, more frequently refreshed rate limits for Google Workspace AI Ultra for Business subscribers.
Google · Gemini CLI·Dec 18, 2025
cherry-pick a6d1245 to release/v0.21.1-pr-15214 to patch version v0.21.1 and create version 0.21.2.
Anthropic · Claude apps·Dec 18, 2025Announcement
In October, we introduced skills—a way to teach Claude repeatable workflows tailored to how you work. Skills are now easier to deploy, discover, and build with organization-wide management for Team and Enterprise plans, a directory of partner-built skills, and an open standard (Agent Skills) so skills work across AI platforms. Claude in Chrome is now available in beta to all paid plan subscribers, including Pro, Team, and Enterprise plans.
OpenAI · New model·Dec 18, 2025Announcement
Today we are releasing GPT-5.2-Codex, the most advanced agentic coding model yet for complex, real-world software engineering. GPT-5.2-Codex is a version of GPT-5.2 further optimized for agentic coding in Codex, including improvements on long-horizon work through context compaction, stronger performance on large code changes like refactors and migrations, improved performance in Windows environments, and significantly stronger cybersecurity capabilities. Starting today, the CLI and IDE Extension will default to gpt-5.2-codex for users who are signed in with ChatGPT. API access for the model will come soon. If you have a model specified in your config.toml configuration file, you can instead try out gpt-5.2-codex for a new Codex CLI session using: bash codex --model gpt-5.2-codex. You can also use the /model slash command in the CLI. In the Codex IDE Extension you can select GPT-5.2-Codex from the dropdown menu. If you want to switch for all sessions, you can change your default model to gpt-5.2-codex by updating your config.toml configuration file: toml model = "gpt-5.2-codex”.
Mistral AI · Mistral API·Dec 18, 2025
Introducing extract_footer, extract_header in our OCR API, as well as hyperlinks output.
Mistral AI · New model·Dec 18, 2025
We released OCR 3 (mistral-ocr-2512).
Google · Gemini CLI·Dec 17, 2025
Google · Gemini API·Dec 17, 2025
Launched Gemini 3 Flash Preview, gemini-3-flash-preview, delivering fast frontier-class performance that rivals larger models at a fraction of the cost. With upgraded visual and spatial reasoning, and agentic coding capabilities. Read the documentation on some new features, including: Multimodal function responses Code execution with images
Google · Gemini app·Dec 17, 2025
What: Gemini 3 Flash is now the new default model in the Gemini app. It offers next generation intelligence at lightning speed and represents a major capability upgrade over Gemini 2.5 Flash. This model provides PhD-level reasoning comparable to larger models and delivers a significant leap in multimodal understanding, allowing you to ask questions using images, audio, text, and more. Gemini 3 Flash can be accessed in the Gemini app by selecting “Fast” for quick answers and “Thinking” to help solve more complex problems faster. Gemini 3 Pro is also available as “Pro” in the model picker and remains the best choice for advanced math and coding. Why: We want to make Gemini the most helpful AI assistant for everyone. By upgrading our default model to Gemini 3 Flash, we’re bringing next generation intelligence to all your tasks, helping you tackle your most ambitious projects, faster.
Google · Antigravity·Dec 17, 2025
Support for Gemini 3 Flash in Antigravity.
Google · Gemini CLI·Dec 16, 2025
enable message bus integration by default. consolidate remote MCP servers to use url in config.
Google · Gemini app·Dec 16, 2025
What: We're rolling out new experimental Gems from Google Labs. With this feature, you can create custom AI mini-apps to help you with everyday tasks or explore new ideas. Try a premade Gem made by Labs like ‘Recipe Genie’ that helps transform your leftovers into delicious meals or ‘Claymation Expert’ that turns any topic into a claymation infographic. Remix to make it your own or start from scratch to bring your ideas to life. Why: We want to make Gems even more helpful to meet your custom needs.
Mistral AI · New model·Dec 16, 2025
We released Mistral Small Creative (labs-mistral-small-creative) as a Labs model.
Google · Gemini CLI·Dec 12, 2025
cherry-pick edbe548 to release/v0.20.1-pr-15007 [CONFLICTS].
cherry-pick 3f5f030 to release/v0.20.0-pr-14843 to patch version v0.20.0 and create version 0.20.1.
Google · Gemini API·Dec 12, 2025
Released gemini-2.5-flash-native-audio-preview-12-2025, a new native audio model for the Live API. This update improves the model's ability to handle complex workflows. To learn more, see the Live API guide and Gemini 2.5 Flash Native Audio.
Google · Gemini API·Dec 11, 2025
Launched the Interactions API. This API provides a unified interface for interacting with Gemini models and agents. To learn more, see the Interactions API guide. Launched the Gemini Deep Research agent in preview. It can autonomously plan, execute, and synthesize results for multi-step research tasks.
Google · Gemini API·Dec 10, 2025
Launched enhancements to our text-to-speech models, Gemini 2.5 Flash TTS preview (optimized for low latency) and Gemini 2.5 Pro TTS preview (optimized for quality), including enhanced expressivity, precision pacing, and seamless dialogue.
Google · Gemini CLI·Dec 9, 2025
Improve request token calculation accuracy. Hook Tool Execution Integration.
Google · Gemini API·Dec 9, 2025
The following Gemini Live API models are now shut down: gemini-2.0-flash-live-001 gemini-live-2.5-flash-preview
Mistral AI · Mistral API·Dec 9, 2025
We released Mistral Vibe.