Devstral 2 (123B)

The license on Devstral 2 (123B) is the single most important thing to understand before evaluating this model. Mistral's announcement described it as "modified MIT," but as developers on X correctly flagged, the practical effect is closer to a proprietary license with MIT-style attribution: commercial use above a revenue threshold requires a separate paid agreement with Mistral. This is a fundamentally different license posture from Devstral Small 2 (24B), which is clean Apache 2.0. If you're choosing between the two for a commercial product, Devstral Small 2 is almost always the correct default — the license clarity is worth more than the 4-percentage-point SWE-Bench difference. Use Devstral 2 (123B) when the additional capability genuinely matters AND your deployment either falls below the revenue threshold OR you have a commercial license. On capability: this is genuinely the strongest open-weightA model where the trained weights are freely downloadable — you can run it yourself without contacting the creator. Llama, Mistral, Qwen, and Gemma are open-weight. Open-weight does not mean open-source: the training data and code often stay private. The license still governs what you can do with the weights, including whether you can use them commercially. agentic coding model available. On infrastructure: the 123B dense architecture needs a multi-GPUThe specialized chip that runs most AI models. Originally designed for 3D graphics, GPUs turned out to be excellent at the math AI requires. Nvidia dominates the AI GPU market; common datacenter models include the H100, H200, and B200. Running an AI model without a GPU is possible but painfully slow for anything but the smallest models. node (not laptop-feasible).