garas (Gemma 4 31B)

plantwatererDecision modelOpen weights

garas (Gemma 4 31B) by plantwaterer is a decision model: it answers typed questions with a probability for each allowed answer instead of generating text. It scores 58.3 on the Jev Decision Index (rank 18).

Pricing

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

Specifications

Provider: plantwaterer
Parameters: 31B
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: google/gemma-4-31B-it
Jev-compatible request: No
Runs on CPU: No
License: apache-2.0
Hugging Face: plantwaterer/garas

Benchmarks

Jev Decision IndexECE 0.049: 58.3

Jev Decision Index: public 56.6, ECE 0.049, Brier 0.369. 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: 56.6

Speed & latency

Median latency per call (Decision Index)

117 ms

Compare garas (Gemma 4 31B) with…

garas (Gemma 4 31B) vs amos-1-large—/1M in garas (Gemma 4 31B) vs amos-1-medium—/1M in garas (Gemma 4 31B) vs Celeris-1 Decision$0.04/1M in garas (Gemma 4 31B) vs ScaleDown Classify-1$0.04/1M in garas (Gemma 4 31B) vs Cloudflare clef—/1M in garas (Gemma 4 31B) vs Cloudflare clef-flash—/1M in

FAQ

How much does garas (Gemma 4 31B) cost?

garas (Gemma 4 31B) has no hosted price listed; its weights are open, so the cost is the compute you run it on.

What questions can garas (Gemma 4 31B) answer?

garas (Gemma 4 31B) answers choice (one of n options), score (ordered rubric), yes / no questions, returning a probability for each allowed answer.

Is garas (Gemma 4 31B) well calibrated?

On the Jev Decision Index its expected calibration error (ECE) is 0.049; below about 0.05, a stated 80% confidence is right roughly 80% of the time.

Is garas (Gemma 4 31B) open weight, and can it run on a CPU?

Yes, garas (Gemma 4 31B) has open weights. It needs a GPU to run at useful speed.

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.