Devstral Small (24B)

Qwen 2.5

Qwen 2.5–72B is Alibaba’s flagship open-weight 72B dense LLM for high-accuracy reasoning, coding, and multilingual tasks. It supports long-context prompting, tool/function calling, and reliable JSON outputs—ideal for enterprise RAG, agents, and repo-scale analysis.

Magistral Small 1.2

Magistral Small 1.2 is Mistral AI’s compact reasoning model, optimized for efficiency and cost-sensitive deployments. With strong instruction following, reliable reasoning, and structured JSON outputs, it’s suited for lightweight copilots, chat assistants, and automation workflows where speed and affordability are key.

Mistral Small 3