SimpleJev
Featherless AI's open server that turns a stock Hugging Face language model into a typed-decision endpoint. It reads the next-token logits for each question and builds the Choice, Score or Noul answer itself; no classifier head is trained and no JSON is generated.
There are no SimpleJev weights: the server runs models such as Gemma-4-26B-A4B (the public demo), Qwen3.8-27B or Qwen3.5-0.8B, plus a native backend for Laya. Up to 255 options per Choice, up to 256 questions per request, and a configurable context. A free demo API (2k tokens, 2 requests per second) needs no key; higher limits are on Featherless's paid plans. Featherless publishes no accuracy of its own and says it does not reproduce Jev's training. Third-party results: JevBench v1.4.2 ranks the Qwen3.8-27B demo 34th (86.6% public, 35.7% sealed); Decision Index 0.2.1 scores it 55.74. Server code Apache-2.0; the served model's own licence applies.
What it decides
- choice — picks one option from a set
- score — places the input on an ordered scale
- noul — answers a yes/no question with one calibrated probability
- classify — assigns a category from a fixed taxonomy
- route — sends the input to one of several destinations
At a glance
| Base model | google/gemma-4-26b-a4b-it |
| Maker | Featherless AI |
| Released | 2026-09-18 |
| License | apache-2.0 |
Hosted API
Served by Featherless AI. Get access.
Read more
This page was opened by System One for Featherless AI, who can claim the organisation and take it over at any time.