# Decision Index 0.3 Retrieval

> Decision Index 0.3 publishes chance-corrected retrieval scores for 112 evaluated systems.

Canonical page: https://benchlm.ai/benchmarks/jevdecisionindex03retrieval

- Category: [Decision Models](/decision-models)
- Last updated: October 6, 2026

The Decision Index 0.3 board lists 112 evaluated systems. This snapshot follows source commit f03e5a6 and keeps the retrieval score values and size labels published on the board.

The full score weights 37 public benchmarks at 20%, private tests of the same skills at 50%, and private tasks from new domains at 30%. The source reports 110,201 decisions. Scores are chance-corrected and coverage-adjusted; unsupported or unanswered requests count as wrong.

The published aggregates include private task results, but the task contents and item-level answers are not public. We show the board as external reference data, separate from general model rankings.

- [Decision Index 0.3 board](https://huggingface.co/spaces/multimodalart/jev-decision-index)
- [Decision Index methodology](https://huggingface.co/spaces/multimodalart/jev-decision-index/blob/main/methodology.html)
- [Space README and model repositories](https://huggingface.co/spaces/multimodalart/jev-decision-index/blob/main/README.md)
- [Decision Models](/decision-models)

## About Decision Index 0.3 Retrieval

- Year: 2026
- Tasks: Retrieval area of the Decision Index 0.3 evaluation
- Format: Chance-corrected area score
- Difficulty: Retrieval and classification decision tasks
- Paper: [Decision Index 0.3 leaderboard](https://huggingface.co/spaces/multimodalart/jev-decision-index)

The area score appears on the board beside the full score. The full score combines public benchmarks at 20%, private tests of the same skills at 50%, and private tasks from new domains at 30%. Private task contents and item-level results are not published.

Decision Index 0.3 Retrieval is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (112 evaluated systems)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Perplexity Decider v1.1 (27B)](/models/decision-index-0-3-perplexity-decider-v1-1-27b) | Perplexity | 64.7 points |
| 2 | [GEV-26B-Decide](/models/decision-index-0-3-gev-26b-decide) | autotrust | 64.6 points |
| 3 | [Torchcast Decision 27B](/models/decision-index-0-3-torchcast-decision-27b) | Not listed | 64.5 points |
| 4 | [deck31b (Gemma 4 31B)](/models/decision-index-0-3-deck31b-gemma-4-31b) | Not listed | 64.2 points |
| 5 | [Quyet-1.0-Large](/models/decision-index-0-3-quyet-1-0-large) | chinhnc | 64.0 points |
| 6 | [simple-jev · Qwen3.8-27B (featherless)](/models/decision-index-simple-jev-qwen3-8-27b) | Featherless AI | 63.4 points |
| 7 | [Decider chat · Gemma-4-31B](/models/decision-index-decider-chat-gemma4-31b) | Mapika | 63.3 points |
| 8 | [Blink v0.3 26B-A4B (NVFP4)](/models/decision-index-0-3-blink-v0-3-26b-a4b-nvfp4) | PixilabAI | 63.1 points |
| 9 | [Surogate Rune 26B-A4B v3](/models/decision-index-rune-26b-a4b-v3) | Surogate | 62.7 points |
| 10 | [Jev](/models/decision-index-0-3-jev) | Not listed | 61.4 points |
| 11 | [Xor 26B-A4B](/models/decision-index-0-3-xor-26b-a4b) | juspay | 61.1 points |
| 12 | [vLLM-SR Decision 2.0 Vega 27B](/models/decision-index-0-3-vllm-sr-decision-2-0-vega-27b) | Not listed | 61.0 points |
| 13 | [Cloudflare clef](/models/cloudflare-clef) | Cloudflare | 60.7 points |
| 14 | [Fastino GLiDE no-thinking (28B)](/models/decision-index-0-3-fastino-glide-no-thinking-28b) | Not listed | 60.0 points |
| 15 | [Eikos-27B](/models/eikos-27b) | caiovicentino1 | 58.3 points |
| 16 | [reflex 27B](/models/reflex-27b) | kshetrajna12 | 58.1 points |
| 17 | [JEV-27B](/models/decision-index-0-3-jev-27b) | autotrust | 57.1 points |
| 18 | [Xor 1.2](/models/decision-index-0-3-xor-1-2) | Not listed | 56.9 points |
| 19 | [decisio 0.4.0 (Gemma 4 12B)](/models/decision-index-0-3-decisio-0-4-0-gemma-4-12b) | Not listed | 56.9 points |
| 20 | [Jebadiah 27B](/models/decision-index-0-3-jebadiah-27b) | frontier-infra | 55.8 points |
| 21 | [Kev 27B](/models/decision-index-0-3-kev-27b) | jaredpalmer | 55.6 points |
| 22 | [JADE](/models/decision-index-0-3-jade) | Not listed | 55.5 points |
| 23 | [JoshuaSP diffusiongemma (open-jev)](/models/decision-index-0-3-joshuasp-diffusiongemma-open-jev) | Not listed | 55.4 points |
| 24 | [AJev lora5 (Gemma 4 12B)](/models/decision-index-0-3-ajev-lora5-gemma-4-12b) | andyzhang232 | 55.3 points |
| 25 | [JPT-35B-A3B](/models/decision-index-0-3-jpt-35b-a3b) | kirp | 55.2 points |
| 26 | [vLLM-SR Decision 2.0 Lux 9B](/models/decision-index-0-3-vllm-sr-decision-2-0-lux-9b) | Not listed | 54.7 points |
| 27 | [Winnow-12B](/models/winnow-12b) | Eldan Ring | 54.4 points |
| 28 | [Jevfire](/models/decision-index-0-3-jevfire) | Not listed | 54.4 points |
| 29 | [decisio 0.4.0 (Qwen3.6-35B-A3B)](/models/decision-index-0-3-decisio-0-4-0-qwen3-6-35b-a3b) | Not listed | 53.9 points |
| 30 | [Bespoke Nimble 9B v3](/models/decision-index-0-3-bespoke-nimble-9b-v3) | bespokelabs | 53.6 points |
| 31 | [Decider chat · Qwen3.6-27B](/models/decision-index-0-3-decider-chat-qwen3-6-27b) | Not listed | 53.3 points |
| 32 | [ezjev 4B s2](/models/decision-index-0-3-ezjev-4b-s2) | everettjf | 53.2 points |
| 33 | [Hopper 12B (trained)](/models/decision-index-0-3-hopper-12b-trained) | Not listed | 53.0 points |
| 34 | [Decider 35B-A3B](/models/decider-35b-a3b) | Mapika | 52.9 points |
| 35 | [vLLM-SR Decision 2.0 Nox 4B](/models/decision-index-0-3-vllm-sr-decision-2-0-nox-4b) | Not listed | 52.3 points |
| 36 | [JevK5 9B v0.3.3](/models/decision-index-0-3-jevk5-9b-v0-3-3) | alibiserikbay | 51.5 points |
| 37 | [jiwo 4B](/models/decision-index-0-3-jiwo-4b) | eljiwo | 51.3 points |
| 38 | [Jev-Style 2B v3](/models/decision-index-0-3-jev-style-2b-v3) | Not listed | 50.1 points |
| 39 | [JEV-9B](/models/decision-index-0-3-jev-9b) | autotrust | 47.9 points |
| 40 | [Jet v6.2](/models/decision-index-0-3-jet-v6-2) | michaljach | 47.8 points |
| 41 | [Kev 9B v2](/models/decision-index-0-3-kev-9b-v2) | Not listed | 47.3 points |
| 42 | [RSI-Jev v5.0 VL 3B](/models/decision-index-0-3-rsi-jev-v5-0-vl-3b) | shgao | 46.8 points |
| 43 | [Ryotide Qwen v0.3.1](/models/decision-index-0-3-ryotide-qwen-v0-3-1) | Not listed | 46.3 points |
| 44 | [Hopper (G) 1.2](/models/decision-index-0-3-hopper-g-1-2) | HopitAI | 46.2 points |
| 45 | [Sieve 9B](/models/decision-index-0-3-sieve-9b) | sthanika-ai | 45.9 points |
| 46 | [JevK5 v0.3](/models/decision-index-0-3-jevk5-v0-3) | Not listed | 45.3 points |
| 47 | [Lev1 3.7B](/models/decision-index-0-3-lev1-3-7b) | Not listed | 44.9 points |
| 48 | [NeoHorse-Jev-4B](/models/decision-index-0-3-neohorse-jev-4b) | TokenRhythm | 44.4 points |
| 49 | [Cloudflare clef-flash](/models/cloudflare-clef-flash) | Cloudflare | 44.1 points |
| 50 | [djev](/models/djev) | Maisa | 44.0 points |
| 51 | [lev](/models/decision-index-0-3-lev) | interfaze-ai | 43.9 points |
| 52 | [JPT-9B](/models/decision-index-0-3-jpt-9b) | kirp | 43.0 points |
| 53 | [jiwo 0.8B](/models/decision-index-0-3-jiwo-0-8b) | eljiwo | 43.0 points |
| 54 | [Decider 4B](/models/decision-index-0-3-decider-4b) | Mapika | 42.9 points |
| 55 | [Kev 4B r10](/models/decision-index-0-3-kev-4b-r10) | jaredpalmer | 41.9 points |
| 56 | [vLLM-SR Decision 2.0 Sol 2B](/models/decision-index-0-3-vllm-sr-decision-2-0-sol-2b) | Not listed | 41.9 points |
| 57 | [JPT-4B](/models/decision-index-0-3-jpt-4b) | kirp | 41.7 points |
| 58 | [Winnow-E4B](/models/decision-index-0-3-winnow-e4b) | EldanRing | 41.6 points |
| 59 | [Kas 4B](/models/decision-index-0-3-kas-4b) | kpiya | 40.7 points |
| 60 | [Intern-Decision-4B](/models/decision-index-0-3-intern-decision-4b) | internlm | 39.8 points |
| 61 | [Bosun v3.1 1.7B](/models/decision-index-0-3-bosun-v3-1-1-7b) | Hanno-Labs | 39.3 points |
| 62 | [Bekko System One v0 400M](/models/decision-index-0-3-bekko-system-one-v0-400m) | hotchpotch | 39.0 points |
| 63 | [Gevva E2B](/models/decision-index-0-3-gevva-e2b) | davidburhans | 38.5 points |
| 64 | [Jev-Omni](/models/jev-omni) | akhilaaa3 | 38.3 points |
| 65 | [razorback16 openjev diffusiongemma (NVFP4, vLLM)](/models/decision-index-0-3-razorback16-openjev-diffusiongemma-nvfp4-vllm) | Not listed | 37.5 points |
| 66 | [Intelif Qwen3-4B](/models/decision-index-0-3-intelif-qwen3-4b) | Not listed | 37.5 points |
| 67 | [Jevstral 8B](/models/decision-index-0-3-jevstral-8b) | AmirBraham | 37.4 points |
| 68 | [Jobe](/models/decision-index-0-3-jobe) | Not listed | 37.3 points |
| 69 | [Decider 2B](/models/decider-2b) | Mapika | 36.9 points |
| 70 | [this-that 1.2](/models/decision-index-0-3-this-that-1-2) | Not listed | 36.4 points |
| 71 | [Bespoke Nimble 9B v2](/models/decision-index-0-3-bespoke-nimble-9b-v2) | bespokelabs | 35.7 points |
| 72 | [Gevva E4B](/models/decision-index-0-3-gevva-e4b) | davidburhans | 34.9 points |
| 73 | [vLLM-SR Decision 2.0 Eos 0.8B](/models/decision-index-0-3-vllm-sr-decision-2-0-eos-0-8b) | Not listed | 34.6 points |
| 74 | [EXAONE-4.0-1.2B-JEV v0.3](/models/decision-index-0-3-exaone-4-0-1-2b-jev-v0-3) | Not listed | 34.0 points |
| 75 | [Dinah-0](/models/decision-index-0-3-dinah-0) | Lukitaduarte | 33.6 points |
| 76 | [Kodiak v0.2 1B (accuracy)](/models/decision-index-0-3-kodiak-v0-2-1b-accuracy) | cortex-agent-llc | 31.7 points |
| 77 | [Sifr 0.8B v3.1](/models/decision-index-0-3-sifr-0-8b-v3-1) | Not listed | 31.6 points |
| 78 | [OneJev 0.8B](/models/decision-index-0-3-onejev-0-8b) | OmniJev | 30.9 points |
| 79 | [Bosun v3.1 0.6B](/models/bosun-v31-0-6b) | Clause Logic | 29.9 points |
| 80 | [Bekko System One v0 68M](/models/decision-index-0-3-bekko-system-one-v0-68m) | hotchpotch | 29.6 points |
| 81 | [openvons](/models/decision-index-0-3-openvons) | Not listed | 28.7 points |
| 82 | [vLLM-SR Decision 2.0 Kai 0.6B](/models/decision-index-0-3-vllm-sr-decision-2-0-kai-0-6b) | Not listed | 28.2 points |
| 83 | [Sieve 2B](/models/decision-index-0-3-sieve-2b) | sthanika-ai | 28.1 points |
| 84 | [Solomon v1.1](/models/decision-index-0-3-solomon-v1-1) | Not listed | 26.6 points |
| 85 | [open-jev (pngwn)](/models/decision-index-0-3-open-jev-pngwn) | Not listed | 25.4 points |
| 86 | [Von 1.3](/models/decision-index-0-3-von-1-3) | Not listed | 23.5 points |
| 87 | [mmastrac diffusiongemma (vLLM PR 57250)](/models/decision-index-0-3-mmastrac-diffusiongemma-vllm-pr-57250) | Not listed | 23.4 points |
| 88 | [Jev-Style 0.8B v3](/models/decision-index-0-3-jev-style-0-8b-v3) | Not listed | 23.0 points |
| 89 | [JevK5 Lite](/models/decision-index-0-3-jevk5-lite) | alibiserikbay | 22.7 points |
| 90 | [GLiNER2.5-Decide](/models/decision-index-0-3-gliner2-5-decide) | fastino | 20.3 points |
| 91 | [jeff](/models/jeff) | Logan Markewich | 18.4 points |
| 92 | [Kev 0.8B r15](/models/decision-index-0-3-kev-0-8b-r15) | Not listed | 18.3 points |
| 93 | [Tev1-4B-experimental](/models/decision-index-0-3-tev1-4b-experimental) | togethercomputer | 17.6 points |
| 94 | [SemIf](/models/decision-index-0-3-semif) | Not listed | 17.2 points |
| 95 | [Intern-Decision-2B](/models/decision-index-0-3-intern-decision-2b) | internlm | 17.1 points |
| 96 | [Bekko System One v0 17M](/models/decision-index-0-3-bekko-system-one-v0-17m) | hotchpotch | 16.4 points |
| 97 | [Lavoir](/models/decision-index-0-3-lavoir) | moganai | 15.6 points |
| 98 | [JPT-0.8B](/models/decision-index-0-3-jpt-0-8b) | kirp | 15.1 points |
| 99 | [Metask-Jev-4B](/models/metask-jev-4b) | Wayfind (metask-ai) | 13.8 points |
| 100 | [DM-JEPA 1.1](/models/decision-index-0-3-dm-jepa-1-1) | Not listed | 12.8 points |
| 101 | [system-one-gemma](/models/decision-index-0-3-system-one-gemma) | Not listed | 11.2 points |
| 102 | [mini-jev](/models/decision-index-0-3-mini-jev) | Not listed | 9.6 points |
| 103 | [Laya](/models/laya) | Convai Innovations | 8.6 points |
| 104 | [Intern-Decision-0.8B](/models/decision-index-0-3-intern-decision-0-8b) | internlm | 7.9 points |
| 105 | [CLM-v0.1-8B](/models/decision-index-0-3-clm-v0-1-8b) | Contrastive-LM | 6.5 points |
| 106 | [LFM2.5-2.6B-RLCD](/models/decision-index-0-3-lfm2-5-2-6b-rlcd) | monotykamary | 6.5 points |
| 107 | [Julia 1](/models/decision-index-0-3-julia-1) | SupersonicLabs | 5.3 points |
| 108 | [Tev1-0.8B-experimental](/models/decision-index-0-3-tev1-0-8b-experimental) | togethercomputer | 3.8 points |
| 109 | [Qwen-2.5-1B-RLCD](/models/decision-index-0-3-qwen-2-5-1b-rlcd) | harshatheg | 1.0 points |
| 110 | [Verdict](/models/decision-index-0-3-verdict) | Not listed | 0.0 points |
| 111 | [LFM2.5-350M-RLCD](/models/decision-index-0-3-lfm2-5-350m-rlcd) | notnotsamuel | 0.0 points |
| 112 | [Zev](/models/decision-index-0-3-zev) | Not listed | 0.0 points |

## FAQ

### What does Decision Index 0.3 Retrieval measure?

Decision Index 0.3 publishes chance-corrected retrieval scores for 112 evaluated systems.

### Which model leads the published Decision Index 0.3 Retrieval snapshot?

Perplexity Decider v1.1 (27B) currently leads the published Decision Index 0.3 Retrieval snapshot with a score of 64.7 points.

### How many models are evaluated on Decision Index 0.3 Retrieval?

The October 6, 2026 contains 112 AI models.

### Does Decision Index 0.3 Retrieval affect BenchLM's overall score?

Not directly. Decision Index 0.3 Retrieval is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.