# Decisio 12B tuned v2 commit decision model by tachara-ai > Decisio 12B tuned v2 commit returns typed option probabilities through decisio. Read how its yes/no commit rule changes calibration and serving. Updated 2026-11-07 (Australia/Adelaide) ## Quick answers ### What is Decisio 12B tuned v2 commit by tachara-ai? [Decisio 12B tuned v2 commit](): A Gemma 4 12B fine-tune with merged weights for the decisio runtime. [Decisio 12B tuned v2 commit facts]() [Source]() ### Who makes Decisio 12B tuned v2 commit? [Decisio 12B tuned v2 commit](): tachara-ai [Decisio 12B tuned v2 commit facts]() [Source]() ### Are Decisio 12B tuned v2 commit's weights open? [Decisio 12B tuned v2 commit](): A Gemma 4 12B fine-tune with merged weights for the decisio runtime. Licence: Apache-2.0, per card and included licence [Decisio 12B tuned v2 commit facts]() [Source]() By tachara-ai · Weights upload: 2026-11-06 · OPEN A Gemma 4 12B fine-tune with merged weights for the decisio runtime. Its commit rule pushes yes/no probabilities inside the 0.20 to 0.80 band to that band’s edge on the same side. This preserves the winning answer while changing calibration. ## Facts - Base and size: Gemma 4 12B; merged bf16 weights - Inputs: State and typed choice, yes/no or score questions - Runtime: decisio 0.35.0 or later; plain vLLM does not apply the full readout - Licence: Apache-2.0, per card and included licence - Access: Public weight shards and decision\_config.json ## Claims and sources - **VENDOR-REPORTED**: The maker says temperatures were fitted by log loss and training was checked against both boards’ public test sets. Training data are undisclosed; the card supplies no task-accuracy result. [source]() ## Links - [Source: Model card and weights]() ## Related projects No related projects are listed in the registry.