Liquid AI's first decision model, served only through the Liquid API. Returns Choice, Score and Noul answers with probability distributions and zero output tokens, on the /v1/systemone request shape used by TypeSafe's SDK.
A Gemma 4 12B fine-tune for typed decisions that also chats and reads images. Ships as GGUF for llama.cpp, holds a 64K context on a 16 GB GPU, and serves /v1/systemone next to /v1/chat/completions.
Decides
choice, score, noul, classify, route
choice, score, noul, classify, route
Architecture
d1
winnow
Fine-tuned from
—
google/gemma-4-12b-it
License
proprietary
apache-2.0
Availability
Hosted API
Open weights
Hosted by
Liquid AI
—
Input price
—
—
Decision accuracy
—
85.7%
Calibration error
—
—
Valid action rate
—
—
Median latency
—
143 ms
p95 latency
—
—
Evaluation suite
Figures are from each model’s manifest; accuracy and latency are what the publishers report, on their own suites and hardware. Add a third model.
Questions
What is the difference between d1 and winnow?
d1 is from Liquid AI and winnow from EldanRing. d1 is only available as a hosted API; winnow has open weights you can download and run. Both answer choice, score, noul, classify and route questions. d1 is licensed proprietary; winnow, apache-2.0.
Which is more accurate, d1 or winnow?
Only winnow publishes an accuracy figure (85.7% on JevBench public subset (231 items), Q8_0); d1 does not, so there is no comparison to make without your own test.
Which is cheaper, d1 or winnow?
d1: Hosted, price not published. winnow: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run d1 or winnow locally?
winnow yes — systemone pull eldanring/winnow downloads its weights. The other is only served as a hosted API.
