Mistral releases Vibe 2.0 terminal coding agent with custom subagents and Devstral 2 API pricing
Mistral AI released Vibe 2.0, a terminal-native coding agent powered by Devstral 2, adding custom subagents, multi-choice clarifications, and slash-command skills. Devstral 2 API pricing is now $0.40/M input tokens and $2.00/M output tokens, with a smaller variant at $0.10/$0.30 per million tokens.
Mistral releases Vibe 2.0 terminal coding agent with custom subagents and Devstral 2 API pricing
Mistral AI released Vibe 2.0, a terminal-native coding agent powered by the Devstral 2 model family. The update adds custom subagents, multi-choice clarifications, slash-command skills, unified agent modes, and automatic CLI updates.
Devstral 2 pricing
Devstral 2 has moved to paid API access with the following pricing:
- Devstral 2: $0.40/M input tokens, $2.00/M output tokens
- Devstral 2 Small: $0.10/M input tokens, $0.30/M output tokens
Free API usage remains available on Mistral's Experiment plan for testing and prototyping.
New features in Vibe 2.0
Custom subagents: Developers can build specialized agents for specific tasks like deployment scripts, pull request reviews, or test generation, then invoke them on demand.
Multi-choice clarifications: When intent is ambiguous, Vibe prompts users with options instead of guessing, requiring confirmation before executing potentially destructive actions.
Slash-command skills: Preconfigured workflows for common tasks like deploying, linting, or generating documentation can be loaded with / commands.
Unified agent modes: Custom modes combine tools, permissions, and behaviors, allowing developers to switch contexts without changing tools.
Automatic updates: Bug fixes and improvements ship continuously to the Vibe CLI without requiring manual updates.
Availability and pricing
Mistral Vibe 2.0 is available on Le Chat Pro and Team plans. According to Mistral, subscribers get "generous usage for full-time development" and can continue beyond limits with pay-as-you-go at API rates. Students receive 50% off. Users can also bring their own API key (BYOK).
The Team plan includes everything in Pro plus unified billing, management, and priority support.
Enterprise services
Mistral offers enterprise add-ons including fine-tuning on internal languages and domain-specific languages (DSLs), reinforcement learning with custom environments, and end-to-end code modernization for migrating entire codebases to modern stacks. The company states it already delivers these solutions to organizations in finance, defense, and infrastructure.
What this means
Mistral is positioning Vibe as a customizable alternative to coding assistants like GitHub Copilot and Cursor, emphasizing control and configurability. The move to paid Devstral 2 API access signals Mistral's shift from pure research releases to monetizing its specialized coding models. The custom subagent feature could differentiate Vibe by allowing teams to encode institutional knowledge and workflows directly into their development tools, though adoption will depend on whether the customization overhead justifies switching from established coding assistants.
Related Articles
Mozilla Launches Firefox Smart Window Assistant Powered by Mistral AI
Mozilla has launched Firefox Smart Window, a beta AI browsing assistant powered by Mistral's models. The feature handles complex searches, tab summarization, and browsing history recall, with a stated commitment to zero data retention.
Mistral AI Models Now Power Mozilla's Firefox Smart Window Browsing Assistant
Mistral AI will power Mozilla's Firefox Smart Window (beta), a browsing assistant that summarizes searches and tracks tabs. The rollout starts in France and North America, with the UK and Germany to follow later this year.
Meta Launches WhatsApp Business Tools MCP to Let AI Agents Automate Business Setup
Meta released a new MCP server that connects AI coding agents directly to the WhatsApp Business Platform, automating account creation, phone verification, and messaging template setup. The move expands Meta's existing lineup of MCP servers beyond ad management and app monitoring tools.
AWS Details How Amazon Bedrock Prompt Caching Cuts Input Token Costs by Up to 90%
Amazon Bedrock's prompt caching feature can cut input token costs by up to 90% on cache hits by storing repeated context like documents, system prompts, and tool definitions. AWS outlines six implementation patterns and pricing details, including a 25% premium for cache writes and 90% discount on cache reads.
Comments
Loading...