Introducing PorTAL: Portable Task Adapters for LLMs. A novel recipe to cheaply port fine-tuning between models. It matches per task LoRA accuracy at half the cost, lowering the switching overhead of adapting tasks across LLMs. At Ramp, every new model release used to mean retrai…↗

Today we’re releasing Ramp SWE-Bench: a private, production-grounded coding benchmark created from real engineering problems we've faced at Ramp.↗

We partnered with @PrimeIntellect to build Fast Ask, a small RL-trained subagent that helps our Sheets agent find answers in spreadsheets. It scores +4% over Opus on exact match accuracy at Haiku latency.↗

x.com/i/article/1957…↗