Postman names Baseten its recommended inference partner for enterprise agents
Overview
Postman names Baseten its recommended inference partner for enterprise teams running models behind Postman-built agents, according to Baseten's own blog post.
Baseten says its Model APIs rank among the top two providers on Artificial Analysis for throughput and latency across seven 2026 launches, and that it retains no data by default.
Baseten also says its Training service post-trains open models for specific agent tasks. It cites Harvey, which reports that a 27B open model matched closed frontier models on its legal agent benchmark, where top closed models cost roughly $50 per task. These are Baseten's and Harvey's claims; the post does not present independent verification of the performance rankings.
Written by AI from the articles below · updated Oct 29, 2:28 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- Baseten BlogOfficialBaseten and Postman pair agent infrastructure with enterprise API tooling
AIPostman names Baseten its recommended inference partner for enterprise teams, which runs models behind Postman-built agents. Baseten says its Model APIs rank among the top two providers on Artificial Analysis for throughput and latency across seven 2026 launches, and it offers zero data retention by default. Baseten Training post-trains open models for specific agent tasks, and Harvey says a 27B open model matched closed frontier models on its legal agent benchmark, where top closed models cost roughly $50 per task.
Heat Trend
Not enough continuous observations to show a trend yet.