Skip to main content

LLM Model Recommendations

We've tested 43 models across major providers against our MCP tools. Two stand out for reliability and performance.

Mistral Medium 3.1​

Our default recommendation. 100% tool-calling accuracy, strong answer quality (0.67 avg score), and fast responses (~6.1s avg latency). Best balance of quality, speed, and cost for most deployments.

Gemini 3 Flash​

Best for high-volume or cost-sensitive deployments. 100% tool-calling accuracy with the lowest price point among reliable models. Slightly higher latency (~11.2s avg) but significantly cheaper per query.