SID-2: an RL-trained agentic search model with a 12-agent swarm mode
SID AI's second-generation search model searches, reads results and re-queries on its own; a 12-agent swarm scores 0.720 NDCG on the company's benchmark.
Why it matters
A single agent answers in 3.1s at 0.585 NDCG while the 12-agent swarm tops GPT-6 Astra's self-reported score — search as an agentic loop, not a fixed pipeline.
What you could build with it
A team building a support product could replace a fixed search pipeline with an agentic search model that re-queries and re-ranks on its own, so customers get better answers from the same document set without reindexing.
Does it hold up?
Too early to judge — every NDCG figure is self-reported by SID AI, with no independent evals and no disclosed model size or weights.
Built with SID-2
- SID AI launches SID-2 search model with 0.720 NDCG in swarm modeoossa · Second writeup of the launch: single-agent 0.585 NDCG in 3.1s, swarm 0.720, pricing from $0.25 per million input tokens.
- To navigate vast amounts of data, SID-2 bets on collective intelligencelhc.media · Independent read of the SID-2 technical report: how the swarm splits the search space, the 506-question benchmark behind the 0.720 NDCG claim, and the caveat that the times are estimates, not production measurements.
- sidhq/sid-gogithub · Official Go SDK for the SID API, for calling SID-2 from Go services.
Learn more
First spotted on article: source.
More AI models
Edge0-145B: open 145B MoE that runs on a single consumer GPUEdge0 open-sourced 145B and 120B latent-MoE models derived from Qwen3.8-Flash-Next, plus an inference framework that…models · JEV 0.74Strata: run a 125B model on a gaming PCFree, MIT-licensed local inference engine that runs Qwen3.8-Flash-Next (125B MoE, about 6B active per token) on…models · JEV 0.73Qwen4-Max: Alibaba's new flagship model goes liveAlibaba's 2.4T-parameter flagship is now generally available via API, with a 1M-token context, native vision, and…models · JEV 0.67smolmaxxing: run 16B LLMs on a 1 GB TV box or a 2013 iPadAn MIT-licensed llama.cpp patch set that runs language models up to a 16B MoE on devices with about 1 GB of usable RAM…models · JEV 0.65Blue TTS v3: a 388MB open text-to-speech model with voice cloningAn open-source 388MB text-to-speech model (MIT) with voice cloning from about 10 seconds of speech, 7 languages, 15…models · JEV 0.65Haru-4BOpen full-duplex spoken-dialogue foundation model that listens and speaks simultaneously (7.5 Hz frames, 133 ms per…models · JEV 0.63
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.