Baseten releases open decision models bison-1 and groundhog-1 on Gemma 4

Nov 21, 2026
Baseten says bison-1 tops the Artificial Analysis Decision Index; both models return probabilities instead of text and have open weights.

Baseten has released bison-1 and groundhog-1, two open-weight "decision models" built on Google's Gemma 4. Decision models do not write text. They return calibrated probabilities for yes/no, multiple-choice and score questions, for tasks such as ticket triage, spam detection and routing requests between models.

bison-1 is a 31B model. groundhog-1 is a cheaper mixture-of-experts model with about 4B active parameters. According to Baseten, bison-1 ranks first of 19 models on the Artificial Analysis Decision Index with a score of 76, ahead of Liquid AI's d1.2-ultra at 75 and Baseten's own Qwen-based mammoth-1 at 71. It also has the lowest Brier score on the index (0.14), a measure of how well stated confidence matches accuracy.

Both models take text and images and offer 128k context, eight times Baseten's previous decision models. They cost $0.10 and $0.035 per million tokens on Baseten's Model APIs, and the weights are on Hugging Face. Baseten says it moved to Gemma for teams that want an American base model.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Daily AI Brief.

Also, consider following us on social media:

Free newsletter

Daily AI Brief

Daily report covering major AI developments and industry news, with both top stories and complete market updates

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more
Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.