Pareto's multi-model routing matches Astra in agent evaluation

elvis @omarsar0Originally posted 2026.09.25

Open original on X

Interesting results here. This is why I expect more agent workloads to run on blended models. Pareto 26.9 from @TheUnbiasedCo sends requests to several frontier and open models and keeps the best answer. In the new eval of 30 agent tasks, Pareto tied GPT-6 Astra for first place at about 1/3 the cost per successful task. It also finished tasks faster than DeepSeek V4 Pro and GLM 5.3 Flash.

This post quotes another post: “Composio tests 6 models on 30 agent tasks”