Mistral releases Mistral Code, a “vibe coding” client forked from open-source project Continue, in private beta on JetBrains development platforms and VS Code
It also follows Mistral's agent API launch, which broadened the company's tooling for code execution and other agent tasks. The client is a practical distribution layer for that broader developer stack, while retaining an open-source starting point in Continue.
First-order effects
Developers admitted to the private beta can use Mistral Code inside JetBrains platforms and VS Code rather than accessing Mistral only through underlying models or APIs.
Mistral becomes a direct product competitor to coding-client vendors such as Windsurf, with a client forked from Continue as its initial implementation base.
Second-order effects
Mistral now has to compete on the day-to-day IDE experience—workflow integration, reliability, and developer adoption—not solely on the capabilities of its coding models.
Continue gains visibility as a foundation for commercial AI coding clients, while rival clients face another model provider distributing through the same major developer environments.
Third-order effects
If model providers continue shipping their own IDE clients, differentiation in AI coding may shift from standalone model access toward control of the developer work surface and its integrations.
The use of open-source client foundations could make the interface layer more contestable, while concentrating competition around proprietary models, agent services, and enterprise deployment choices.
The trend: AI model companies are moving up the stack from supplying coding models and APIs to owning the agentic interfaces where developers actually write and ship software.
Excited to introduce @MistralAI Mistral Code! 🤖 The most customizable AI-powered coding assistant for enterprises! 🛠️ Frontier coding models, fully customizable & tunable to your codebase 🌟 State-of-the-art coding assistance & agentic coding, in your control 🔧 One platform, [imag…
We're rolling out a few updates to Codex today: 1. Codex is rolling out to ChatGPT Plus users today. It includes generous usage limits for a limited time, but during periods of high demand, we might set rate limits for Plus users so that Codex remains widely available.
codex gets access to the internet today! it is off by default and there are complex tradeoffs; people should read about the risks carefully and use when it makes sense. also, we are making in available in the chatgpt plus tier.
also, today we are making a lightweight version of memory available to the free tier of chatgpt! memory has probably become my favorite feature in chatgpt; excited for us to improve this a lot over time.
Codex's internet access comes with a very clear security disclaimer. What's interesting: They even provide an example of prompt injection as part of the docs. I think that's a great move. Educating users is key. [image]
3. Internet access is off by default, and can be enabled when creating a new environment or by editing an existing one. You have full control over the domains and HTTP methods Codex can use during task execution. Learn more about usage and risks in the docs: [image]
2. Next, our top requested feature: You can now give Codex access to the internet during task execution to install base dependencies, run tests that need external resources, upgrade or install packages needed to build new features, and more.
Things are about to get really interesting! 👀 Codex, OpenAI's software engineering agent that can work on many tasks in parallel, is live today for Plus users. [image]
This is a huge upgrade! Imo codex is one of the best coding tools out there today and my main issue with it was its lack of internet access. I wrote about my experience with codex here, it should be even more incredible now: https://x.com/...