Blogs›Mistral AI›Mistral Vibe Terminal Agent
Mistral Vibe now supports remote agents running in the cloud, powered by Mistral Medium 3.5. This enables parallel, asynchronous task execution for coding and productivity work. A new 'Work mode' in Le Chat allows for complex, multi-step tasks involving cross-tool workflows, research, and inbox triage, with human oversight for sensitive actions. Mistral Medium 3.5 is a 128B dense model with a 256k context window, optimized for long-horizon tasks, instruction-following, reasoning, and coding. Le Chat also introduces Deep Research mode for structured reports, Voice mode powered by Voxtral, Projects for organizing conversations, and advanced image editing capabilities.
5/28/2026
This post announces that Vibe is now a unified agent for long-horizon productivity, capable of handling tasks like coding and research. It also mentions that Vibe runs on flagship Mistral models optimized for reasoning, agentic tasks, tool calls, and coding.
5/28/2026
This post announces the rebranding of Le Chat to Vibe, unifying work and coding capabilities into a single agent. It details the new Work Mode for complex, multi-stage tasks across enterprise tools (Google Workspace, Outlook, SharePoint, Slack, GitHub, etc.), including structured data analysis, document synthesis, multi-step task scheduling, and reusable skills. Code Mode is introduced with a web surface for remote coding sessions, integrations with GitHub, and a new VS Code extension. The Vibe CLI is updated with more session controls, skills as slash commands, custom modes/subagents, editable plans, multi-choice questions, session-scoped permissions, and a `/teleport` command. The post also outlines Vibe's pricing plans and availability.
5/22/2026
Introduces Mistral Medium 3.5, a new 128B dense model with a 256k context window, optimized for long-horizon coding and productivity tasks. This model powers the new remote agents in Mistral Vibe, allowing for cloud-based, parallel, and asynchronous task execution. It also enables a new 'Work mode' in Le Chat for complex, multi-step agentic tasks involving cross-tool workflows and research. The post details the model's performance on SWE-Bench Verified and τ³-Telecom benchmarks, and its availability for self-hosting.
1/27/2026
Introduces Mistral Vibe 2.0 with custom subagents, multi-choice clarifications, slash-command skills, and unified agent modes. Details pricing for Devstral 2 API access and outlines enterprise services for fine-tuning, reinforcement learning, and code modernization. Provides installation instructions for the Vibe CLI.
12/9/2025
Introduces Devstral 2 (123B) and Devstral Small 2 (24B), new open-weight coding models. Devstral 2 achieves 72.2% on SWE-bench Verified and is highly cost-efficient. Devstral Small 2 is compact, locally deployable, and supports multimodal agents. Also introduces Mistral Vibe CLI, an open-source command-line coding assistant powered by Devstral, featuring project-aware context, multi-file orchestration, and IDE integration.
7/17/2025
This post introduces several new features for Le Chat, enhancing its capabilities for research, voice interaction, reasoning, organization, and image editing. Deep Research mode leverages a tool-augmented agent for structured, source-backed reports. Voice mode is powered by the new Voxtral model for low-latency speech recognition. Reasoning capabilities are enhanced by the Magistral model for multilingual and complex thought processes. Projects feature allows users to organize conversations into context-rich folders. Advanced image editing is integrated in partnership with Black Forest Labs.