User repository on GitHub
An open System One decision model
- Forks
- 3
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Models & Reimplementations
Lev — abhinavexists
An open System One decision model from Interfaze-ai: give lev a state — text, a ticket, an email, or JSON — plus yes/no, choice, or score questions and it answers all of them in one forward pass from logits, returning calibrated probabilities and zero output tokens.
- Stars
- 56
- Language
- Python
- License
- APACHE-2.0
- Reading time
- 3 min
- Published by
- Community
- Author
- Abhinavexists
- Last verified
- Oct 2026
Used inJev for Software & DevTools
Highlights
- README: lev is an open System One decision model and the harness that measures it — one state, many typed decisions, one forward pass, no generated tokens.
- It returns calibrated probabilities over exactly the options you supplied; the model is a LoRA adapter on Qwen3.5-4B, trained on one H100.
- 74.3% on all 13 S1Bench subsets with 4B parameters; on the six subsets the public board completed it sits behind only Jev and two open models of 26B and up.
- The wire protocol is TypeSafe's /v1/systemone, so any TypeSafe client works against it by changing the base URL.
- Apache-2.0 Python repo: 56 stars, 3 forks, created 2026-09-20, pushed 2026-10-15; weights at huggingface.co/interfaze-ai/lev.
Watch out
README reviewed 2026-10-17 and truncated inside the navigation links, so quickstart commands, serving hardware, and the full benchmark harness are not in the digest. The 74.3% score and the placement relative to other models are the project's own claims against the public S1Bench board; no independent replication appears in the excerpt.
Questions about this entry
Short answers drawn from the linked project or page.
An open System One decision model from Interfaze-ai: you give it a state — text, a ticket, an email, or JSON — and typed questions such as yes/no, choice, and score, and it answers all of them in one forward pass, returning probabilities while generating nothing.
It is its own model, a LoRA adapter on Qwen3.5-4B with weights on the Hugging Face Hub, but it speaks TypeSafe's /v1/systemone wire protocol, so any TypeSafe client works against it by changing the base URL. The launch post is titled "Jev now open source — lev".
The README headline is 74.3% across all 13 S1Bench subsets with 4B parameters and zero output tokens; on the six subsets the public board had completed, it trailed only Jev and two open models of 26B and up, ahead of reflex-4b and a 35B model.
Training was a Qwen3.5-4B LoRA on one H100. The digest truncates before the quickstart section, so serving requirements, quantization options, and latency numbers are not available in the excerpt — the README does link a speed receipt in docs/DECISIONS.md.
Point any TypeSafe client at a server speaking the /v1/systemone protocol with lev loaded — that is the integration the README documents — but the digest truncates before the Quickstart anchor, so the exact serve command is not available here.
Interfaze-ai: the README links a blog post titled "Jev now open source — lev" at interfaze.ai, weights at huggingface.co/interfaze-ai/lev, and design documents including FINDINGS.md and DECISIONS.md. CI runs on GitHub Actions.
A benchmark harness to measure the model yourself, a decision log including ADR-023's choice of one batched forward over prefill-and-fork, a snake demo showing decisions in action, and a train-it-yourself path; the README's navigation is truncated in the digest.
Other listings of this project
Forks and copies of this project in the directory. This entry is the canonical one.
Open Jev-style decision model built on LiquidAI's LFM2.5-350M backbone, a clone of kev. Serves the System One API contract so the TypeSafe SDK works against a local server, returning probabilities for typed questions from one forward pass.
Jev Models & Reimplementations#open-models#python#local
Communityfranckverrot/lev
More Jev resources like this
A Chinese-language decision model that returns a probability for each fixed candidate option in one forward pass, with no generated text or chain of thought. It is a LoRA fine-tune of Qwen3.5-0.8B on 13,048,000 CJ-Dataset rows, benchmarked on CJ-Bench against Chinese-Jev.
Jev Models & Reimplementations#community#github#python
Community
yanqiangmiffy/chinese-decisionPython
yanqiangmiffy/chinese-decisionPythonA small, fast decision model — a system one model — for agentic workflows from Strands Labs: it picks between options or rates on a scale faster than an LLM and returns a calibrated confidence with every decision.
Jev Models & Reimplementations#community#github#python
Community
strands-labs/strands-deciderPython
strands-labs/strands-deciderPythonSmall Jev-style decision models that read a state and typed questions and return a probability for every option in one forward pass without generating text; the 0.75B and 4.2B versions speak TypeSafe's Jev API format at a median of about 50 ms per request on one H100.
Jev Models & Reimplementations#community#github#decision-model
Community
jiwidi/jiwoPython
jiwidi/jiwoPythonSystem One models from first principles by Qi Li and Xinchao Wang at the National University of Singapore: give Jeva a state and typed questions and it returns candidate probabilities, a Boolean probability, or an expected rubric score, with reusable option memory.
Jev Models & Reimplementations#community#github#python
Community
liqiiiii/jevaPython
liqiiiii/jevaPythonAn open-source, local alternative to Jev: System-One decision models that answer the POST /v1/systemone shape — choice, yes/no, or score — for general decisions and browser-agent steps. Models ship as OpenJev-1.7B, 4B, and 8B, with a pip-installable Python package.
Jev Models & Reimplementations#community#github#python
Community
alanhuangyoo/openjevPython
alanhuangyoo/openjevPythonA System One layout optimizer for decision-model workloads: it reorders rows and fields so a prefix-caching backend reuses input computation, then returns per-record decisions in original order. Its bundled backend serves AutoTrust JEV-9B through vLLM.
Jev Repos & SDKs#community#github#python
Community
0814wdwd/solo_jevPython
0814wdwd/solo_jevPythonFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.