Ecosystem Timeline

Gemini 4 Hydrogen is generally available to Copilot Pro, Pro+, Max, Business, and Enterprise customers, with gradual rollout across supported Copilot clients. It is billed at provider list pricing; Business and Enterprise admins manage access through model policy.

The Copilot app now generally supports GitHub Enterprise Server 3.20+ with version 1.2.0 or later, including remote repositories, issues, and pull requests. Cloud sessions are not available for GitHub Enterprise Server.

GitHub plans to retire Claude Fable 5, Claude Haiku 4.5, Gemini 3.8 Flash, Gemini 4 Argon, Grok 4.6, and Kimi K3 on 2026-12-14. Suggested successors include Fable 5.2, Haiku 5.5, Gemini 4 Hydrogen, Gemini 4 Carbon, and Grok 4.7; GitHub has not announced a Kimi K3 replacement.

Google releases Gemini 4 Hydrogen (gemini-4-hydrogen) for fast daily coding and agent workflows. The stable model supports multimodal input, tool use, and a 1,048,576-token unified context window. Google recommends it as the migration target for Gemini 3.8 Flash.

Google releases Gemini 4 Hydrogen (gemini-4-hydrogen), a fast model for daily coding and agent workflows. It has a 1,048,576-token unified context window, supports multimodal input and tool use, and is the recommended migration target for Gemini 3.8 Flash. The Live API is not supported.

Gemini 4 Carbon becomes available to paying API and business developers. Introductory pricing is $2 input / $10 output per 1M tokens through 2026-11-30, then $4 / $20.

Gemini 4 Carbon is generally available for Copilot Pro+, Max, Business, and Enterprise across supported Copilot clients, with provider list pricing and Business/Enterprise model-policy controls.

GitHub Apps can use the enterprise Copilot seat-management API with the Enterprise Copilot license management fine-grained permission.

Anthropic releases Fable 5.2 at $8 input / $40 output per 1M tokens. Mythos 5.2 is the same underlying model for verified cyber and life-sciences programs.

Sarvam appointed Sashikumar Sreedharan to lead enterprise, government, and commercial operations as the company expands its full-stack AI offering.

Fable 5.2 becomes broadly available at $8 input / $40 output per 1M tokens. Mythos 5.2 is the same underlying model for Anthropic’s Cyber and Life Sciences Verification Programs.

Claude Fable 5.2 is generally available for Copilot Pro+, Max, Business, and Enterprise across supported Copilot clients. Anthropic safety-classifier retention applies by default; eligible enterprises can use zero data retention.

The @GitHub experience in Slack and Teams is generally available for investigating bugs, creating issues, requesting code changes, and opening pull requests from conversations.

OpenAI releases gpt-6-cyber, a cyber-specialized GPT-6 model for authorized vulnerability research. Access is initially limited to approved Daybreak Red organizations through the API and Codex.

OpenAI announces that it intends to acquire Pierre Computer Company, describing the transaction as an expansion of Codex tools for building, understanding, and reviewing code. The announcement does not state that the transaction has closed.

OpenAI releases Images 2.6 in ChatGPT, Codex, and the API. The API adds GPT-Image-2.6 Flare and Sunburst, with complexity controls, markup-guided editing, SVG export, and animated GIF generation.

GitHub Copilot is available in public preview through Xcode’s built-in coding assistant beginning with Xcode 27.3 beta. The existing Copilot for Xcode extension remains a separate option.

Snowflake announces CoCo and CoWork updates focused on governed business context and agentic action, alongside broader Cortex AI platform releases for enterprise deployment.

GitHub Copilot becomes available in JetBrains Air for local coding tasks and, for Air Teams-enabled organizations, isolated cloud tasks. The integration uses the Agent Client Protocol; existing Copilot subscriptions apply.

Mistral publishes improved Mistral Large 4 weights at mistralai/Mistral-Large-4-1T-A52B. The API id remains mistral-large-4; docs also show version pin mistral-large-4+1. The model card still labels the API Public Preview. The license is the Mistral Ecosystem Open-Weight License, a modified MIT: use, modification, fine-tuning, and commercialization are allowed, except a Model-as-a-Service business needs a separate commercial license. API price is $0.70 input / $0.07 cached input / $2.10 output per 1M tokens. Reasoning efforts are low, medium, and high (default).

Cursor removed Anthropic.claude-code 2.1.312 and 2.1.313 for macOS and Windows from the Cursor VS Code Extension Marketplace. Those builds were listed from about 2:27am to 1:35pm PT and were not published by Anthropic. Cursor shut the marketplace, rotated keys, and brought it back. The notice says other versions and the official Claude Code distribution were not identified as affected. Uninstalling the extension does not clear every change those builds made; follow the remediation on the notice if you installed either build in that window.

Google announces Gemini 4 Carbon for cybersecurity and long-horizon work. Access starts with a subset of cyber defenders, then additional Fairwind partners. Developer, business, and consumer access is planned later. Google has not published a public API id. Introductory price is $2 input / $10 output per 1M tokens, with cached input 95% off, then $4 / $20 after an introductory period whose end date is not stated. Argon remains the public Gemini 4 API id.

Paid Copilot plans can retain personal preferences and repository conventions across supported Copilot clients. Memory is a context feature, not a substitute for repository access controls or review.

Copilot Fusion is generally available in VS Code, the Copilot app, Copilot CLI, and JetBrains. Standard, High, and Max set effort. Copilot routes across enabled models with one model, a cascade, or a draft-review-revise pass. Billing is each model’s standard rate, with a limited-time 30% discount on paid plans. Earlier auto, HydraFusion, and HydraFusion Max selections move to the matching Fusion levels. Administrators can set a Fusion effort level as the default.

GitHub Agentic Workflows, from GitHub Next, is generally available on GitHub Free, Team, and Enterprise. Natural-language Markdown compiles to GitHub Actions YAML and runs GitHub Copilot or another coding agent. The default Copilot CLI engine requires a Copilot license. Workflows are read-only by default.

Paid Copilot plans gain one resumable cloud session across GitHub.com, GitHub Mobile, the Copilot app, CLI, VS Code, Slack, and Teams. Sessions run in GitHub-managed cloud sandboxes by default; organization policy, repository exclusions, and MCP allowlists still apply.

Copilot can generate or edit images from chat in VS Code, the Copilot app, and Copilot CLI. The preview is available to Pro, Pro+, Business, and Enterprise plans; image tokens are billed separately, and organization owners can disable the model through policy.

Projects in the GitHub Copilot app give multi-repository work one place for context, agents, and results. You describe the goal, attach repositories and other context, and Copilot runs as many agents as the work needs. Public preview.

Agent Merge watches a pull request and fixes selected blockers in a cloud development environment: failing Actions workflows, merge conflicts, or both. It is in the Copilot app, VS Code, and the merge box on github.com. Paid Copilot plans, subject to organization and enterprise policy. It can be combined with auto-merge.

DeepMind announces Genie 4, a real-time world model with native audio and language steering, for trusted testers. Google AI Ultra access in Project Genie, and early access for Gemini API and AI Studio partners, are described as coming later. No public model id or price. It is not a Gemini chat model.