Sthanika Sieve 31B: cost-efficient decision model family for agents
Hyderabad-based applied-AI lab Sthanika (PurpleTalk group) announced Sieve 31B, a decision model ranked second globally on the open Jev Decision Index with a full score of 63.8, alongside 2B+/4B/9B+ siblings.
Why it matters
A small decision model built to replace general LLMs for thousands of routine agent choices — returning a decision plus confidence score at a fraction of inference cost.
What you could build with it
A startup building multi-agent support or RPA bots could route every routine classification and policy decision through these tiny models, cutting inference spend by an order of magnitude while reserving the big model for hard cases.
Does it hold up?
Too early to judge — the claim rests on a company press release distributed via NewsVoir/ANI; the benchmark methodology and scores have not been independently verified.
Built with Sthanika Sieve 31B
- Business Fortnight: Sthanika Sieve 31B ranks second on Jev Decision Indexblog · Second-source coverage of the announcement with performance details across the Sieve family size classes.
- Sahyadri Startups: Indian AI lab Sthanika Sieve 31B beats Jev, Perplexity on decision benchmarkblog · Regional startup-press coverage of the announcement, repeating the 63.8 full score and the family-size leaderboard placements.
Learn more
First spotted on article: source.
More AI models
OpenAI publishes 722 AI-generated math manuscripts from an unreleased modelOpenAI released 722 mathematical manuscripts in 372 result families to GitHub under Apache-2.0, produced by an…models · JEV 0.74StepFun Step-5-Preview — open BF16 weights releasedStepFun published the full BF16 weights for its flagship Step-5-Preview agentic model (600B-parameter sparse MoE, 27B…models · JEV 0.73ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72Mistral Large 4 (Le Chonk): 1.05T-param open-weight MoEMistral's new flagship: a 1.05T-parameter mixture-of-experts multimodal model with 49B active params and 1M-token…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71JetBrains Mellum 2.1: 12B MoE coding model trained for agentsJetBrains released Mellum 2.1, a 12B MoE model (2.5B active parameters, 128K context) under Apache 2.0, trained mainly…models · JEV 0.69
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.