Beating GPT-5.6 Sol on retrieval with 100x cheaper open models
The article introduces Castform, a tool designed to help developers post-train smaller, open-source models to perform retrieval tasks as effectively as frontier models like GPT-5.6. By optimizing these models, developers can achieve significant cost reductions in agentic search workflows.
Why it matters
Reducing the cost and latency of AI agents is a major hurdle for widespread enterprise adoption, and this approach offers a more efficient alternative to relying solely on expensive proprietary models.
Back to Blog / Product How Castform + Neon Beats Frontier Models on Price and Efficiency A 4B open-source model post-trained with Castform retrieved search results as accurately as GPT-5.6 Sol, while costing 100x less
Receive only our latest updates. No spam, guaranteed.
“Most teams' best training data is just sitting in their databases. The problem is that turning raw data into something usable is hard, and letting agents read, search, and mutate data cheaply at scale requires advanced infra. Pointing Castform at Neon skips both.”
A "good agent" needs to be strong in 2 areas:
Neon (Lakebase Postgres) and their new Search extensions solve the first; Castform solves the second.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in