ENPT

oenogarum

Last updated:

Open259–259ms$0/M input

Quadrant scores

See the full quadrant

Scored with the public quadrant rubric (maturity × capability; bubble = adoption). Revise as evidence lands.

  • Maturity4.8/10

    Apache-2.0 full fine-tune of google/gemma-4-31B-it (31.3B, bf16) with a calibrated last-layer head. Needs the repo's own attention function and Decision Index engines. GitHub mirror holds code. No hosted API or SLA.

  • Capability6.7/10

    Decision Index 0.4.1 (29 Oct) Full 68.10 (#11 of 247), ECE 0.040, median 259 ms. Choice and yes/no only. Published method: temperature T = 1.080 plus a Laplace last-layer correction on yes/no.

  • Adoption1/100

    About 199 downloads, 0 likes, 0 GitHub stars (30 Oct). No integrations or production users found.

Vendor claims

publicado (OK do Fabricio 30/10 BRT). Scores are calculated from scratch.

What it is

oenogarum is a fine-tune of google/gemma-4-31B-it by plantwaterer, the author of garum. It answers multiple-choice and yes/no decision questions with a probability per option. Source: Hugging Face · GitHub mirror.

Board result

Board Row Result
Decision Index 0.4.1 (data 29 Oct) oenogarum Full 68.10 (#11 of 247), public 62.80; ECE 0.040; median 259 ms

Not found on JevBench or the Artificial Analysis Decision Index on 30 Oct; the boards are never combined.

Vendor claims (UnverifiedClaim)

  • Calibrated last-layer head: choice probabilities are the letter softmax divided by T = 1.080; yes/no gets a Bayesian (Laplace) last-layer residual (source)
  • Length-adaptive option-order averaging for requests up to L = 939 characters (source)

Scores (from scratch, 30 Oct 2026)

  • Capability 6.7 = types × 0.30 + calibration × 0.25 + latency × 0.25 + accuracy × 0.20 = 6 × 0.30 + 7 × 0.25 + 6 × 0.25 + 8 × 0.20 = 1.80 + 1.75 + 1.50 + 1.60 = 6.65 → 6.7. Types 6: choice and yes/no, no score (as garum). Calibration 7: named method on the card plus a board ECE of 0.040. Latency 6: board median 259 ms (p50 ≤ 300 ms step). Accuracy: 8 × 68.10 ÷ 59.22 = 9.20 → capped at 8 (one source, not #1).
  • Maturity 4.8 = availability × 0.30 + docs × 0.25 + integrations × 0.25 + support × 0.20 = 7 × 0.30 + 7 × 0.25 + 2 × 0.25 + 2 × 0.20 = 2.10 + 1.75 + 0.50 + 0.40 = 4.75 → 4.8.
  • Adoption 1 = engagement × 0.35 + ecosystem × 0.35 + production × 0.30 = 1 × 0.35 + 1 × 0.35 + 0 × 0.30 = 0.70 → 1.