openjev

Last updated:

Open30–200ms$0/M input

Quadrant scores

See the full quadrant

Scored with the public quadrant rubric (maturity × capability; bubble = adoption). Revise as evidence lands.

  • Maturity5.5/10

    MIT weights, modeling helpers, SGLang package, HF Space, and a public product domain (openjev.co coming soon); still an independent research release without hosted SLA.

  • Capability6.8/10

    Trained 3-way NLI enables choice via rerank and boolean-like checks; 0.8B long-context v2s plus multimodal 4B v2 (image/video/text; audio flagged as coming) broaden the surface — still not a TypeSafe choice/score/noul wire format.

  • Adoption32/100

    HF likes ~379 (was ~287 on 20 Sep) plus Sam Witteveen walkthrough, game demos, and domain presence; still niche outside the open-Jev wave.

Vendor claims

News covering this model

Board audit of 17 Oct 2026: see the “Board audit” section below; where it conflicts with older board statements on this page, it wins.

Not to be confused with OpenJev 27B (openjev/openjev, CC BY-NC 4.0), the 27B model that llama.cpp added to /v1/systemone on 2 Oct 2026. Same name, different author, license and architecture; this page is the MIT NLI stack by AlexWortega.

openjev is an independent open decision stack — Qwen3.5 fine-tuned as a cross-encoder NLI head (contradiction / entailment / neutral) and framed as a Jev-style “universal classifier” — not a TypeSafe product and not a TypeSafe /v1/systemone clone.

Weights and code: AlexWortega/openjev (MIT). Demo / long-context small checkpoint: qwen3.5-0.8b-nli-v2s-long/ (what the HF Space serves). Recommended multimodal checkpoint: qwen3.5-4b-nli-v2/ (text + images; author also describes video-text training). Also ships a text-only 4B v1, a Qwen3.5-35B-A3B MoE variant, optional latent MLP heads, and an external SGLang classify package under code/sglang_openjev/.

Product presence: openjev.co (“Coming Soon” as of 21 Sep 2026; author thanks @_akhaliq for the domain). HF Space: AlexWortega/openjev.

Typical use: score hypotheses about a shared state (predict / predict_hypotheses), rerank answer candidates, or grade against a reference. Sam Witteveen’s walkthrough (Open Jev Models Are Here!!, 20 Sep 2026) demos visual entailment on a checkout UI plus Doom/Minecraft decision loops.

Author X update (21 Sep 2026, @justALEXWORTEGA): 0.8B Qwen variant, multimodal image–video–text input with audio coming, RLCD calibration note, SGLang support claimed “5× faster than Jev” (UnverifiedClaim), and 4B / 35B + blog “this week” — treat ship dates as forward-looking, not catalog facts.

Limits (read before you ship)

  • Catalog decisionTypes list choice and boolean via entailment/rerank — there is no first-class TypeSafe score / noul wire API here
  • Author NLI / game numbers, RLCD vs GRPO, and SGLang “5× faster than Jev” are self-reported → UnverifiedClaim relative to ModelSystem.One benches; JevBench rankings elsewhere are third-party
  • Different contract than hosted Jev or Decider System One JSON — do not drop-in swap without your own adapter
  • Multimodal path needs the v2 checkpoint and enough VRAM for Qwen3.5-4B; audio is announced, not shipped as fact here

Adjacent specialist (not this model)

anespo28/openjev-fincalc-4b is a MIT LoRA fine-tune of this OpenJev checkpoint as an NLI judge for financial arithmetic (LTV, DSCR, covenants). The author reports 94.4% on a synthetic FinCalc-NLI test set (vs 63.2% for openjev-4B v5) — UnverifiedClaim, synthetic data only. It stays adjacent / specialist: domain-narrow, not a general System One catalog page. See docs/04.

Why it is in the catalog

Public trained decision weights (not frozen-logit method-only), clear non-affiliation, and a distinct NLI/multimodal peer next to Laya / NanoJev. Roundup context: Sam Witteveen.

Board audit (17 Oct 2026)

Checked against the official boards on 17 Oct 2026: Decision Index 0.3.2 (data of 16 Oct, 19:00 UTC), JevBench and ImageJevBench live. The boards are separate sources and are never averaged. Older board statements and rankings further down this page are superseded where they conflict with this section.

Rows with similar names on the boards. The repo links differ from this page’s repo, so they are not attributed to this model:

Board Row Result
Decision Index 0.3.2 (16 Oct 19:00 UTC) open-jev (pngwn) Full 32.09 (#117, tie group 30); public 30.49 (#123); same-skills 30.03; new-domains 34.28
JevBench open weights (live 17 Oct) OpenJev Capability 75.9; Intelligence 70.3; calibration 81.4; p50 0.857 s; $0.24/1k
JevBench open weights (live 17 Oct) Open-Jev 27B v1.1 Capability 68.0; Intelligence 52.5; calibration 83.5; p50 0.790 s; $0.72/1k
JevBench open weights (live 17 Oct) Open-Jev 9B Capability 61.8; Intelligence 43.9; calibration 79.8; p50 0.618 s; $0.17/1k
JevBench open weights (live 17 Oct) Open-Jev 2B Capability 47.7; Intelligence 21.8; calibration 73.6; p50 0.437 s; $0.17/1k

Version / family note. The OpenJev / Open-Jev rows on JevBench and open-jev (pngwn) on Decision Index link to other repositories than this page’s (AlexWortega/openjev). No row is attributed here; score unchanged (6.8) until the repo identity is confirmed.

Hugging Face on 17 Oct (AlexWortega/openjev): 0 downloads, 649 likes, licence tag mit, last modified 2026-10-01.