jev-decision-index / README.md
multimodalart's picture
multimodalart HF Staff
Add new submissions; score notes for some rows
de702e1 verified
|
Raw History Blame Contribute Delete
12.8 kB
metadata
title: Jev Decision Index
emoji: 🔬
colorFrom: yellow
colorTo: blue
sdk: static
header: mini
pinned: false
short_description: Benchmarks and news on various repros of TypeSafe's Jev
thumbnail: >-
  https://huggingface.co/spaces/multimodalart/jev-decision-index/resolve/main/og.png
models:
  - MachinePulseAI/karpo-decision-31B
  - fastino/GLiDE-27B
  - decisionmodels/REAX-1.3
  - logitshq/Logits-Sigma-1-Duo
  - Poindexter-Labs/Copernicus-31B
  - PelaAI/KnowLine-31B-Gen1.5
  - Poindexter-Labs/Copernicus-57B
  - 1000luca/FlexDM-LoRA-v2.1
  - sthanika-ai/Sieve-31B-v3
  - OmniJev/OneJev1.6g-31B
  - turion-ai/turion-jev-31b
  - plantwaterer/eshbar-lulilla
  - tattrongvu/Jeffan-27B-1211
  - vllm-sr/d3-ultra
  - bespokelabs/Nimble-v4-max
  - h2oai/h2o-lightning-31b
  - tattrongvu/Jeffan-31b-1710
  - logitshq/Logits-Sigma-1
  - plantwaterer/liquamen
  - apertixai/cerno-27b-v1.1
  - Rover00/WeMM-JEV
  - baselabs/mammoth-1
  - agentscope-ai/Piccolo-27B
  - Ainvest-World/Ainvest-Decision-27b-v1.0
  - plantwaterer/oenogarum
  - autotrust/AutoTrust-JEV-27B-v1.1
  - startlux-models/StartLux-Decision-v1.1-27B-Instant
  - PNPM244/oak1.1
  - andyzhang232/ajev-gemma4-31b-lm6q
  - ninhnv/chon-1.1-31b
  - PNPM244/decy-preview
  - LycheeAI/Lychee-27B
  - vllm-sr/d3-turbo
  - vllm-sr/d3
  - plantwaterer/allec
  - taoszhang/VisionJEV
  - morriszjm/AnyJev-Tacitus-31B
  - scar-ai/vera-large-2o
  - Mapika/decider-31b
  - PelaAI/KnowLine-35B-A3B-Gen1.5
  - OneShotOrg/OneShot
  - Valen-Team/Valen-Decision-1.5-27B
  - sanghwa-na/bobcat-1.3
  - mukaddamzaid/amos-1-large
  - Maincode/matilda-decision-v1
  - PelaAI/KnowLine-26B-A4B-Gen2.5
  - frontier-infra/jebadiah-27b-v2-1
  - FINAL-Bench/Darwin-27B-RSI-Decision
  - fastino/GLiDE-9B
  - torchcast-ai/torchcast-decision-27b
  - tattrongvu/Jeffan-9B-1211
  - OmniJev/OneJev1.6-27B
  - jsaurabh/qwen-decision-27b-noncausal-lora
  - chinhnc/Quyet-1.0-Large
  - plantwaterer/garum
  - baselabs/elephant-1
  - StandardThinking/StandardOne-35B-SH
  - ZTE-AIM/NTele-Hybrid-27B-v1
  - FrontiersMind/Lumma-fev-27B
  - sthanika-ai/Sieve-27B-v2
  - moganai/mogan-decision-31B-it
  - MISA-core/linne-fern-1-0
  - jaredpalmer/kev-27b
  - juspay/Xor-26B-A4B
  - newsbang/adjudex-9b-v5
  - lihe07/span-02
  - apertixai/cerno-9b-v1.2
  - plantwaterer/garas
  - shisa-ai/shisa-de-2.1
  - caiovicentino1/Eikos-27B-FP8
  - shgao/rsi-jev-v6.1-vl-27b
  - PelaAI/KnowLine-9B-Gen1
  - h2oai/h2o-lightning-bolt-74b
  - startlux-models/StartLux-Decision-v1.1-9B-Instant
  - vllm-sr/Decision-2.0-Vega-27B
  - autotrust/JEV-35B
  - autotrust/JEV-27B
  - agentscope-ai/Piccolo-9B
  - matthewagi/Tern-35B-A3B
  - surogate/rune-26b-a4b-GGUF
  - SkyPanther/synack-decide-26b-a4b
  - PelaAI/KnowLine-4B-Gen5.5
  - PixilabAI/Blink-v0.3-26B-A4B-NVFP4
  - Valen-Team/Valen-Decision-1.5-9B
  - theunnecessarythings/JADE
  - Patrick-18/Dexy-Decision-9B-V2
  - org2ai/Wald-26B-A4B
  - andyzhang232/ajev-gemma4-26b-a4b-lora1
  - Cloudflare/clef
  - startlux-models/StartLux-Decision-v1.1-4B-Instant
  - vllm-sr/d3-flash
  - kirp/jpt-35b-a3b
  - quincyqiang/chinese-decision-pointer-4b
  - apertixai/cerno-4b-v1.2
  - juspay/xor
  - autotrust/GEV-26B-Decide
  - victorbennett/Verdin-4B-v1.2
  - OmniJev/OneJev1.6-9B
  - scar-ai/vera-medium-omni
  - WangSenQWQ/Senova-4B-v4
  - nace-ai/drex-v2.0
  - plantwaterer/muria
  - tattrongvu/Jeffan-4B-2010
  - LatentMesh-Labs/latentmesh-blitz-4b-v1
  - GyrTechnology/saif-4b
  - CountingSheep/vev-4b-v2
  - bespokelabs/Bespoke-Nimble-9B-v3
  - frontier-infra/jebadiah-9b-v2-1
  - agentscope-ai/Piccolo-4B
  - vllm-sr/d3-mini
  - shgao/rsi-jev-v7.0-vl-4b
  - Valen-Team/Valen-Decision-1.5-4B
  - fdemelo/jegue
  - shgao/rsi-jev-v7.0-vl-3b
  - andyzhang232/ajev-gemma4-12b-lora7
  - OmniJev/OneJev1.6-4B
  - onlyaady/BaseDecision-Dyno
  - h2oai/h2o-lightning-e4b
  - sky7350/Mica-v0.5-4B
  - olaverse/PurpleMIST-Flash-1.0
  - sthanika-ai/Sieve-9B-v2
  - Mapika/decider-35b-a3b-nvfp4
  - EldanRing/Winnow-12B
  - ironbcc/Seb-9B-v2
  - olaverse/PurpleMIST-Mini-2.0
  - vllm-sr/Decision-2.0-Lux-9B
  - nace-ai/drex-v1.5
  - HopitAI/hopper
  - kadirbekovvv/jevora-9b-v1
  - corners-ai/CoCo-Decision-2.4-4B
  - Cloudflare/clef-omni
  - alibiserikbay/JevK5-9B
  - Cloudflare/clef-flash
  - eljiwo/jiwo-4b
  - kirp/jpt-9b
  - everettjf/ezjev-4b-s2
  - akhilaaa3/Jev-Omni
  - DoccyHealth/Solomon
  - fastino/GLiDE-2B
  - StandardThinking/StandardOne-8B-SH
  - MLukauskas/amber-hybrid-4b
  - autotrust/JEV-9B
  - bespokelabs/Bespoke-Nimble-9B-v2
  - sthanika-ai/Sieve-9B
  - sthanika-ai/Sieve-4B
  - jaredpalmer/kev-9b
  - blackmallangcow/nari-decision-4b
  - empiriolabsai/aplomb-1
  - vllm-sr/Decision-2.0-Nox-4B
  - HopitAI/hopper-g
  - startlux-models/StartLux-Decision-v1.1-2B-Instant
  - flymy-ai/decision-4b-v1.6
  - StandardThinking/StandardOne-3B-SH
  - kirp/jpt-4b
  - MLukauskas/amber-4b
  - internlm/Intern-Decision-4B
  - PelaAI/KnowLine-2B-Gen1
  - Valen-Team/Valen-Decision-1.5-2B
  - TokenRhythm/NeoHorse-Jev-4B
  - alibiserikbay/JevK5
  - Mapika/decider-4b
  - fastino/GLiDE-0.8B
  - baselabs/joey-1
  - interfaze-ai/lev
  - theaviv/Lev1-3.7B
  - chaoliangUNSW/Jev-Style-2B-Decision-v4.2
  - jaredpalmer/kev-4b
  - chiranko/PKU-KCL-Meteor-2B
  - sthanika-ai/Sieve-2B-v3
  - LatentMesh-Labs/latentmesh-blitz-2b-v1
  - felhen-ai/saracura-ptbr-4b
  - LiquidAI/d1-2B
  - cbuk10001/Cam1
  - vllm-sr/d3-nano
  - kpiya/kas-4b
  - fastino/GLiDE-0.4B
  - togethercomputer/Tev1-4B-experimental
  - LiquidAI/d1-Omni-2B
  - newsbang/adjudex-0.8b-v5
  - EldanRing/Winnow-E4B
  - Leptos-ai/Aurelis-2B
  - knowledgator/gliclass-v4.1-edge
  - AmirBraham/jevstral-8b
  - wayfind/metask-jev-4b-policy-mix
  - UserMoonlight/intelif-qwen3-4b
  - LiquidAI/d1-3B
  - pngwn/system-one-qwen3.5-4b-scorer-v2b
  - chiranko/PKU-KCL-Meteor-0.8B
  - scar-ai/vera-core-2
  - tardellirs/crivo-prumo-1b
  - OmniJev/OneJev1.6-0.8B
  - Kyunghwan-kim/crystal-0.8b
  - newrexlabs/MoDe-L
  - vllm-sr/vela3-fast
  - tardellirs/crivo-1b
  - PelaAI/KnowLine-0.8B-Gen1
  - davidburhans/gevva-e4b
  - startlux-models/StartLux-Decision-v1.1-0.8B-Instant
  - devops-thiago/classone-qwen3.5-4b
  - MLukauskas/amber-0.4b-v2
  - Valen-Team/Valen-Decision-1.5-0.8B
  - run-saina/saina-helm-2-omni-0.8b
  - newrexlabs/MoDe-M
  - GyrTechnology/saif-0.8b-lightning
  - vllm-sr/d3-lite
  - eljiwo/jiwo-0.8b
  - run-saina/saina-helm-2-0.8b
  - tardellirs/crivo-prumo-400m
  - flock-io/this-that-model-1.2
  - MLukauskas/amber-150m
  - newrexlabs/MoDe-S
  - davidburhans/gevva-e2b
  - tardellirs/crivo-400m
  - Mapika/decider-2b
  - vllm-sr/Decision-2.0-Sol-2B
  - AnamK/jdi-ptr14-0.8b
  - vllm-sr/vela3-fast-vision
  - sthanika-ai/sieve-2b
  - teilomillet/wyat-1-800m
  - internlm/Intern-Decision-2B
  - tardellirs/crivo-prumo-68m
  - sthanika-ai/Sieve-0.8B
  - StrandsAgents/strands-decider-2B-hobson-v21
  - carrtesy/EXAONE-4.0-1.2B-JEV
  - tasksource/tasksource-decider-nano
  - tardellirs/crivo-68m
  - MLukauskas/amber-68m
  - devops-thiago/classone-qwen3.5-2b
  - Elfsong/jev-ettin-150m-v16
  - vllm-sr/Decision-2.0-Eos-0.8B
  - tardellirs/crivo-150m
  - hotchpotch/bekko-system-one-v0-400m
  - codavidgarcia/mola-150m
  - vllm-sr/d3-edge
  - codavidgarcia/mola-68m
  - Elfsong/jev-flan-t5-large-20261017-public-v2
  - decision-tune/decisiontune-1.0
  - Hanno-Labs/bosun-v3.1-1.7b
  - vllm-sr/Decision-2.0-Kai-0.6B
  - Lukitaduarte/dinah-1
  - jaredpalmer/kev-0.8b
  - tardellirs/crivo-prumo-32m
  - MLukauskas/amber-32m
  - lCREEl/pick-150m
  - kirp/jpt-0.8b
  - Elfsong/jev-ettin-400m-20261013-v0
  - mohamedlotfy50/sifr-0.8b-v3
  - tardellirs/crivo-32m
  - tardellirs/crivo-prumo-17m
  - tardellirs/prumo
  - cortex-agent-llc/kodiak-v0.2-1b-accuracy
  - internlm/Intern-Decision-0.8B
  - chaoliangUNSW/Jev-Style-0.8B-Decision-v3
  - togethercomputer/Tev1-0.8B-experimental
  - hotchpotch/bekko-system-one-v0-68m
  - MLukauskas/amber-17m
  - LiquidAI/d1-omni-600M
  - tardellirs/crivo-17m
  - wfzyx/von
  - Hanno-Labs/bosun-v3.1-0.6b
  - scar-ai/vera-spark
  - moganai/lavoir
  - fastino/GLiNER2.5-Decide
  - convaiinnovations/laya
  - alibiserikbay/JevK5-Lite
  - monotykamary/LFM2.5-2.6B-RLCD
  - hotchpotch/bekko-system-one-v0-17m
  - Ajeet-001/embeddinggemma2-decision-router
  - Contrastive-LM/CLM-v0.1-8B
  - notnotsamuel/LFM2.5-350M-RLCD
  - DangerLabs/DM-JEPA
  - SupersonicLabs/Julia-1
  - kshitijthakkar/nirnaya-0.4b-0.2a-decision
  - heman10x/rlcd-modernbert-151m
  - shisa-ai/shisa-de-2
  - OmniJev/OneJev1.5-0.8B
  - OmniJev/OneJev1.5-9B
  - OmniJev/OneJev1.5-27B
  - leinenterprises/Herbert_v0.5
  - tinnel123/jev-d3-46k
  - ken-jo/QEV-2B
  - ken-jo/QEV-0.8B
  - sthanika-ai/Sieve-Vision-4B
  - RBDash-Team/RBDash-Decision-v0.2
  - sthanika-ai/Sieve-Vision-27B
  - edward1503/VISTA-4B-A
  - stupid-zwl/JDI-Decision-V1
  - RadixArk/Qwen3.8-27B-NVFP4
  - ukisai/Swift-1.5-Qwen3.8-27b-NVFP4
  - agentscope-ai/Piccolo-27B-VL-1108
  - agentscope-ai/Piccolo-9B-VL
  - agentscope-ai/Piccolo-4B-VL
  - Ashan0525/LingeeJevMulti-4B-v1
  - facebook/meta-encoder-v1
  - shgao/rsi-jev-v5.0-vl-3b
  - AlexWortega/openjev
  - bespokelabs/Bespoke-Nimble-9B
  - C-Tianyu/NanoJev
  - DavidHatley/system-one-mini
  - DeepMostInnovations/sales-conversion-model-reinf-learning
  - fastino/gliner2.5-base-v1
  - fastino/gliner2.5-multi-v1
  - fastino/gliner2.5-small-v1
  - flock-io/this-that-model-1.0
  - frontier-infra/jebadiah-27b
  - FrontiersMind/Lumma-fev-0.1b
  - FrontiersMind/Lumma-fev-0.6b
  - google/diffusiongemma-26B-A4B-it
  - google/gemma-4-31B
  - harshatheg/Qwen-2.5-1B-RLCD
  - knowledgator/gliformer-large-v1
  - llm-semantic-router/Decision-1.0-Eos-0.8B
  - llm-semantic-router/Decision-1.0-Kai-0.6B
  - llm-semantic-router/Decision-1.0-Lex-0.6B
  - llm-semantic-router/Decision-1.0-Lux-9B
  - llm-semantic-router/Decision-1.0-Nox-4B
  - llm-semantic-router/Decision-1.0-Sol-2B
  - Manav2op/verdict-small
  - Mapika/decider-35b-a3b
  - MoLeMo-Lab/mojev
  - perplexity-ai/pplx-decider-v1-27b
  - pngwn/system-one-qwen3.5-4b-scorer
  - Qwen/Qwen2.5-1.5B
  - Qwen/Qwen3-4B-Instruct-2507
  - Qwen/Qwen3.5-4B-Base
  - Qwen/Qwen3.6-27B
  - Qwen/Qwen3.8-27B
  - Qwen/Qwen3.8-27B-FP8
  - andyzhang232/ajev-gemma4-12b-lora5
  - Lukitaduarte/dinah-0
  - OmniJev/OneJev-0.8B
  - perplexity-ai/pplx-decider-v1.1-27b
  - corners-ai/CoCo-Decision-4B
  - frontier-infra/jebadiah-9b-v2
  - Maincode/matilda-jev-v1.5
  - sthanika-ai/Sieve-27B
  - sthanika-ai/Sieve-2B-v2
  - scar-ai/vera-core
  - OmniJev/OneJev1.5-4B

Jev Reproductions Tracker

Who is rebuilding TypeSafe's Jev (System One / RLCD) in the open?

This static Space opens on the Decision Index leaderboard; the News tab tracks the artifacts in one combined grid, color-coded by kind:

  • Decoding: parallel constrained decoding on stock models (inference technique, no new weights)
  • Diffusion: text diffusion models run in a "Jev mode"
  • Trained: Jev-like scoring heads and fine-tunes, weights often on the Hub, promised models listed last
  • Prior art: "this already exists" claims
  • Explainers: architecture speculation, explainers, benchmarks and roundups

Cards sort by a trending score: ♥ likes on X + 5 × GitHub stars + 8 × Hub likes + views ÷ 500, plus a recency credit (up to 8000, halving daily) so a fresh release is not buried under older, louder ones. The credit is scaled by the item's own engagement, so a brand-new entry nobody has reacted to yet cannot climb on its date alone. Cards added within a day of the metrics snapshot get a green new badge. Use the category and "has" chips to filter, or switch the sort.

…plus a section on what is still not in the open: TypeSafe's weights, the RLCD algorithm, and any open model matching Jev's calibration claims.

Hugging Face models and Spaces are first-class artifacts here, not only GitHub repos.

Decision Index (index.html)

The Index tab (switch at the top of every page) is the leaderboard: every open reproduction and Jev, ranked by the Full score: 60% private tests of the same skills and 40% private tasks from new domains, both on the public scale. The public benchmarks are shown next to every model but do not count, so the ranking rests on tests nobody can train on. The switch on the main chart offers Full score, Public only and Your own (set your own part and area weights). Below the chart sit the per-area radars, every benchmark with its public and private scores, and a full page per model (index.html?model=<engine>; ?model=jev is Jev's own report). A Vision board ranks the image-capable entrants the same way.

The public suite is 37 benchmarks in five areas, each chance-corrected so 0 is random guessing and 100 is perfect; an unanswered request counts as wrong. The private tests and the new-domain tasks are never published. The Methodology page has the full rules, every benchmark's subset and baseline, and the changelog.

Contributing

All news data lives in one JS array (ITEMS) near the bottom of news.html (the News tab; the Index tab is index.html). Open a PR on this Space to add or correct an entry. Please include the announcement post on X if there is one, and the Hub / GitHub link.

Engagement metrics are a snapshot (2026-09-24) from public post metadata and will drift.

Credits

Visual identity and outlined Huggies from HF Huggiverse (Chunte/HFBA). Unofficial and community-maintained; not affiliated with TypeSafe AI.