SID AI Launches SID-1 Agentic Search Model, Claims 99% Lower Cost Than GPT-5.1
SID AI launched SID-1, an agentic search model that it said was trained using large-scale reinforcement learning on turbopuffer.
According to the company, the training ran at bursts above 1,000 queries per second on document corpora of more than 10 million documents across thousands of steps. SID AI said SID-1 delivered 1.9 times the recall of RAG plus rerank while running 24 times faster and 99% cheaper than GPT-5.1.
From the sources (2 posts)
@sirupsenwe are seeing more and more RL puffin', here's a model specifically trained for it on tpuf from @SID_AI
@turbopufferSID-1 is an agentic search model by @SID_AI → 1.9x recall over RAG + rerank → 24x faster, 99% cheaper than GPT-5.1 trained using large-scale RL on turbopuffer at 1k+ QPS bursts over 10M+ document corpora across thousands of steps https:/