Report
Fortytwo introduces Engine, an inference API for open-weight models
Fortytwo says Engine starts with Qwen3.8-27B and is private by default, with zero data retention.
TLDR
Fortytwo introduced Engine, an inference API that runs open-weight models through Swarm Inference, starting with Qwen3.8-27B. The company claims scores 7% higher on average than the base model and up to 25% fewer tokens per task. It says Engine has the same per-token price and comparable speed.
Combined views
—
1 Source, first seen ago
