Devstral-2-123B-Instruct-2512 API
About
Build advanced reasoning agents with Devstral-2-123B-Instruct-2512, a specialized AI model optimized for complex logic and chain-of-thought tasks. This model delivers available completely free of charge, native tool calling support, open weights architecture. Access Devstral-2-123B-Instruct-2512 via the Nvidia API with up to 262K output tokens.
Capabilities
Input Modalities
text
Output Modalities
text
Reasoning
Yes
Structured Output
Yes
Tool Use
Yes
WeightsOpen
Limits
Context Window
262K
Tokens
Input Limit
262K
Tokens
Max Output
262K
Tokens
Pricing
Standard (per 1M tokens)
Input
Free
Output
Free
Frequently Asked Questions
Devstral-2-123B-Instruct-2512 by Nvidia costs Free per 1M input tokens and Free per 1M output tokens.
Model Details
ID
mistralai/devstral-2-123b-instruct-2512
Provider
NvidiaFamily
devstral
Release Date
Dec 8, 2025
Knowledge Cutoff
Dec 1, 2025
API Integration
NPM Package
@ai-sdk/openai-compatible
Environment Variables
NVIDIA_API_KEY
API Base URL
https://integrate.api.nvidia.com/v1