Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

Nirnaya 0.4B

kshitijthakkarDecision modelOpen weights

Nirnaya 0.4B by kshitijthakkar is a decision model: it answers typed questions with a probability for each allowed answer instead of generating text. It scores 0.0 on the Jev Decision Index (rank 159).

Pricing

No hosted price is listed. The weights are open: the cost is your own compute.

Specifications

Provider
kshitijthakkar
Parameters
390M
Released
Oct 2026
Open weights
Yes
Question types
Choice (one of N options), Score (ordered rubric), Yes / no
Mechanism
Decoder LM with a decision readout
Backbone
kshitijthakkar/loggenix-moe-0.4B-0.2A-sft-s4
Jev-compatible request
Yes
Runs on CPU
Yes
License
—
Hugging Face
kshitijthakkar/nirnaya-0.4b-0.2a-decision

Benchmarks

Jev Decision IndexECE 0.046
0.0

Jev Decision Index: public 1.6, ECE 0.046, Brier 0.685. Lower ECE means its probabilities can be taken at face value.

Self-reported by the publisher

The model's own claims, often on its own sets or a self-run copy of a public suite. Not independently verified and never ranked alongside measured results.

Jev Decision Index public suite (self-run)Self-reportedDecision Index public balanced skill, edition 0.3 kit (release-v3), self-run
1.6

Speed & latency

Median latency per call (Decision Index)
14 ms

Compare Nirnaya 0.4B with…

Nirnaya 0.4B vs amos-1-large—/1M inNirnaya 0.4B vs amos-1-medium—/1M inNirnaya 0.4B vs Celeris-1 Decision$0.04/1M inNirnaya 0.4B vs ScaleDown Classify-1$0.04/1M inNirnaya 0.4B vs Cloudflare clef—/1M inNirnaya 0.4B vs Cloudflare clef-flash—/1M in

FAQ

How much does Nirnaya 0.4B cost?

Nirnaya 0.4B has no hosted price listed; its weights are open, so the cost is the compute you run it on.

What questions can Nirnaya 0.4B answer?

Nirnaya 0.4B answers choice (one of n options), score (ordered rubric), yes / no questions, returning a probability for each allowed answer.

Is Nirnaya 0.4B well calibrated?

On the Jev Decision Index its expected calibration error (ECE) is 0.046; below about 0.05, a stated 80% confidence is right roughly 80% of the time.

Is Nirnaya 0.4B open weight, and can it run on a CPU?

Yes, Nirnaya 0.4B has open weights. It is small enough to run on a CPU.

Data as of 2026-10-14. Source: Metatext decision-model catalogue, 2026-10-14 (Jev Decision Index, Hugging Face). Pricing is per 1M tokens (USD); confirm with the provider before production use.