Mistral: Mistral Small 4

Mistral Small 4 is a text-and-image input model from Mistral AI with a 262,144-token context window. It supports tool use and reasoning, giving it coverage across agentic and multi-step workflows. Structured output support is unconfirmed, so developers who depend on guaranteed JSON schemas should verify that independently before committing. At $0.15 per million input tokens and $0.60 per million output tokens, the model sits in the budget-to-mid tier on price. Its blended benchmark score of 6.1 is drawn from a single independent benchmark, so the performance picture is thin and should be treated with caution. Teams running high-volume pipelines with image-plus-text inputs, or those needing long-context processing on a modest budget, have reasonable grounds to shortlist it, but anyone prioritizing well-validated capability should wait for broader benchmark coverage before relying on it for critical workloads.