# Decision Index Retrieval → Decision Index Language

Calibration · https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/jevdecisionindex03language/current · data retrieved 2026-11-21

Decision Index Language is estimated from Decision Index Retrieval with a linear curve fitted on 110 models measured on both: y = 0.8571·x + 0.0242, R² = 0.78, cross-validated error 9.5 pp. It is used for 0 estimates.

## Fitted on

| Model | Decision Index Retrieval | Decision Index Language |
|---|---|---|
| [Bosun v3.1 0.6B](https://benchgap.net/model/bosun-v31-0-6b.md) | 28.1% | 5.1% |
| [Clef](https://benchgap.net/model/cloudflare-clef.md) | 59.8% | 47.8% |
| [Clef-Flash](https://benchgap.net/model/cloudflare-clef-flash.md) | 40.8% | 40.5% |
| [Clef-Omni](https://benchgap.net/model/cloudflare-clef-omni.md) | 56.5% | 40.7% |
| [decider-2b](https://benchgap.net/model/decider-2b.md) | 36.7% | 27.4% |
| [decider-35b-a3b](https://benchgap.net/model/decider-35b-a3b.md) | 52.1% | 48.4% |
| [Bekko System One v0 17M](https://benchgap.net/model/decision-index-0-3-bekko-system-one-v0-17m.md) | 14.5% | 0.6% |
| [Bekko System One v0 400M](https://benchgap.net/model/decision-index-0-3-bekko-system-one-v0-400m.md) | 37.7% | 19.6% |
| [Bekko System One v0 68M](https://benchgap.net/model/decision-index-0-3-bekko-system-one-v0-68m.md) | 28.5% | 6.5% |
| [Bespoke Nimble 9B v2](https://benchgap.net/model/decision-index-0-3-bespoke-nimble-9b-v2.md) | 36.0% | 44.4% |
| [Bespoke Nimble 9B v3](https://benchgap.net/model/decision-index-0-3-bespoke-nimble-9b-v3.md) | 52.9% | 54.6% |
| [Blink v0.3 26B-A4B (NVFP4)](https://benchgap.net/model/decision-index-0-3-blink-v0-3-26b-a4b-nvfp4.md) | 63.1% | 59.3% |
| [Bosun v3.1 1.7B](https://benchgap.net/model/decision-index-0-3-bosun-v3-1-1-7b.md) | 37.9% | 14.7% |
| [CLM-v0.1-8B](https://benchgap.net/model/decision-index-0-3-clm-v0-1-8b.md) | 4.6% | 5.7% |
| [Decider 4B](https://benchgap.net/model/decision-index-0-3-decider-4b.md) | 42.2% | 33.9% |
| [Decider chat · Qwen3.6-27B](https://benchgap.net/model/decision-index-0-3-decider-chat-qwen3-6-27b.md) | 53.8% | 52.6% |
| [decisio 0.4.0 (Gemma 4 12B)](https://benchgap.net/model/decision-index-0-3-decisio-0-4-0-gemma-4-12b.md) | 56.5% | 44.0% |
| [decisio 0.4.0 (Qwen3.6-35B-A3B)](https://benchgap.net/model/decision-index-0-3-decisio-0-4-0-qwen3-6-35b-a3b.md) | 54.0% | 50.2% |
| [deck31b (Gemma 4 31B)](https://benchgap.net/model/decision-index-0-3-deck31b-gemma-4-31b.md) | 64.8% | 57.4% |
| [DM-JEPA 1.1](https://benchgap.net/model/decision-index-0-3-dm-jepa-1-1.md) | 4.3% | 2.6% |
| [EXAONE-4.0-1.2B-JEV v0.3](https://benchgap.net/model/decision-index-0-3-exaone-4-0-1-2b-jev-v0-3.md) | 31.2% | 22.2% |
| [ezjev 4B s2](https://benchgap.net/model/decision-index-0-3-ezjev-4b-s2.md) | 52.0% | 47.1% |
| [GEV-26B-Decide](https://benchgap.net/model/decision-index-0-3-gev-26b-decide.md) | 63.3% | 55.9% |
| [Gevva E2B](https://benchgap.net/model/decision-index-0-3-gevva-e2b.md) | 38.5% | 21.2% |
| [Gevva E4B](https://benchgap.net/model/decision-index-0-3-gevva-e4b.md) | 32.9% | 28.3% |
| [GLiNER2.5-Decide](https://benchgap.net/model/decision-index-0-3-gliner2-5-decide.md) | 18.3% | 15.5% |
| [Hopper 12B (trained)](https://benchgap.net/model/decision-index-0-3-hopper-12b-trained.md) | 52.9% | 47.1% |
| [Hopper (G) 1.2](https://benchgap.net/model/decision-index-0-3-hopper-g-1-2.md) | 46.6% | 39.8% |
| [Intelif Qwen3-4B](https://benchgap.net/model/decision-index-0-3-intelif-qwen3-4b.md) | 36.6% | 36.6% |
| [Intern-Decision-0.8B](https://benchgap.net/model/decision-index-0-3-intern-decision-0-8b.md) | 8.5% | 8.4% |
| [Intern-Decision-2B](https://benchgap.net/model/decision-index-0-3-intern-decision-2b.md) | 17.7% | 28.8% |
| [Intern-Decision-4B](https://benchgap.net/model/decision-index-0-3-intern-decision-4b.md) | 39.5% | 41.6% |
| [JADE](https://benchgap.net/model/decision-index-0-3-jade.md) | 56.0% | 56.5% |
| [JEV-27B](https://benchgap.net/model/decision-index-0-3-jev-27b.md) | 58.5% | 51.2% |
| [JEV-9B](https://benchgap.net/model/decision-index-0-3-jev-9b.md) | 48.4% | 40.7% |
| [Jev-Style 0.8B v3](https://benchgap.net/model/decision-index-0-3-jev-style-0-8b-v3.md) | 21.7% | 12.0% |
| [Jevfire](https://benchgap.net/model/decision-index-0-3-jevfire.md) | 53.7% | 44.7% |
| [JevK5 9B v0.3.3](https://benchgap.net/model/decision-index-0-3-jevk5-9b-v0-3-3.md) | 52.0% | 53.7% |
| [JevK5 Lite](https://benchgap.net/model/decision-index-0-3-jevk5-lite.md) | 20.3% | 13.8% |
| [JevK5 v0.3](https://benchgap.net/model/decision-index-0-3-jevk5-v0-3.md) | 45.6% | 46.3% |
| [Jevstral 8B](https://benchgap.net/model/decision-index-0-3-jevstral-8b.md) | 36.6% | 37.3% |
| [Jobe](https://benchgap.net/model/decision-index-0-3-jobe.md) | 37.5% | 36.0% |
| [JoshuaSP diffusiongemma (open-jev)](https://benchgap.net/model/decision-index-0-3-joshuasp-diffusiongemma-open-jev.md) | 54.2% | 51.9% |
| [JPT-0.8B](https://benchgap.net/model/decision-index-0-3-jpt-0-8b.md) | 14.6% | 19.3% |
| [JPT-35B-A3B](https://benchgap.net/model/decision-index-0-3-jpt-35b-a3b.md) | 55.1% | 57.2% |
| [JPT-4B](https://benchgap.net/model/decision-index-0-3-jpt-4b.md) | 40.4% | 41.3% |
| [JPT-9B](https://benchgap.net/model/decision-index-0-3-jpt-9b.md) | 42.3% | 46.5% |
| [Julia 1](https://benchgap.net/model/decision-index-0-3-julia-1.md) | 3.5% | 1.5% |
| [Kas 4B](https://benchgap.net/model/decision-index-0-3-kas-4b.md) | 40.8% | 37.6% |
| [Kev 0.8B r15](https://benchgap.net/model/decision-index-0-3-kev-0-8b-r15.md) | 16.7% | 12.4% |
| [Kev 27B](https://benchgap.net/model/decision-index-0-3-kev-27b.md) | 55.6% | 57.5% |
| [Kev 4B r10](https://benchgap.net/model/decision-index-0-3-kev-4b-r10.md) | 41.7% | 35.6% |
| [Kev 9B v2](https://benchgap.net/model/decision-index-0-3-kev-9b-v2.md) | 47.4% | 44.3% |
| [Kodiak v0.2 1B (accuracy)](https://benchgap.net/model/decision-index-0-3-kodiak-v0-2-1b-accuracy.md) | 29.9% | 10.6% |
| [Lavoir](https://benchgap.net/model/decision-index-0-3-lavoir.md) | 17.1% | 9.2% |
| [lev](https://benchgap.net/model/decision-index-0-3-lev.md) | 42.3% | 29.5% |
| [Lev1 3.7B](https://benchgap.net/model/decision-index-0-3-lev1-3-7b.md) | 44.0% | 41.8% |
| [LFM2.5-2.6B-RLCD](https://benchgap.net/model/decision-index-0-3-lfm2-5-2-6b-rlcd.md) | 4.9% | 5.4% |
| [LFM2.5-350M-RLCD](https://benchgap.net/model/decision-index-0-3-lfm2-5-350m-rlcd.md) | 0.0% | 2.0% |
| [mini-jev](https://benchgap.net/model/decision-index-0-3-mini-jev.md) | 10.0% | 29.1% |
| [mmastrac diffusiongemma (vLLM PR 57250)](https://benchgap.net/model/decision-index-0-3-mmastrac-diffusiongemma-vllm-pr-57250.md) | 23.0% | 47.7% |
| [NeoHorse-Jev-4B](https://benchgap.net/model/decision-index-0-3-neohorse-jev-4b.md) | 44.7% | 34.2% |
| [open-jev (pngwn)](https://benchgap.net/model/decision-index-0-3-open-jev-pngwn.md) | 24.0% | 36.9% |
| [openvons](https://benchgap.net/model/decision-index-0-3-openvons.md) | 30.0% | 31.7% |
| [Quyet-1.0-Large](https://benchgap.net/model/decision-index-0-3-quyet-1-0-large.md) | 63.2% | 58.1% |
| [Qwen-2.5-1B-RLCD](https://benchgap.net/model/decision-index-0-3-qwen-2-5-1b-rlcd.md) | 1.0% | 7.7% |
| [razorback16 openjev diffusiongemma (NVFP4, vLLM)](https://benchgap.net/model/decision-index-0-3-razorback16-openjev-diffusiongemma-nvfp4-vllm.md) | 37.6% | 46.3% |
| [Ryotide Qwen v0.3.1](https://benchgap.net/model/decision-index-0-3-ryotide-qwen-v0-3-1.md) | 45.8% | 33.2% |
| [SemIf](https://benchgap.net/model/decision-index-0-3-semif.md) | 17.8% | 35.9% |
| [Sieve 2B](https://benchgap.net/model/decision-index-0-3-sieve-2b.md) | 28.0% | 21.7% |
| [Sieve 9B](https://benchgap.net/model/decision-index-0-3-sieve-9b.md) | 45.8% | 45.3% |
| [Sifr 0.8B v3.1](https://benchgap.net/model/decision-index-0-3-sifr-0-8b-v3-1.md) | 26.0% | 14.3% |
| [Solomon v1.1](https://benchgap.net/model/decision-index-0-3-solomon-v1-1.md) | 26.9% | 54.2% |
| [system-one-gemma](https://benchgap.net/model/decision-index-0-3-system-one-gemma.md) | 8.8% | 1.7% |
| [Tev1-0.8B-experimental](https://benchgap.net/model/decision-index-0-3-tev1-0-8b-experimental.md) | 3.4% | 14.4% |
| [Tev1-4B-experimental](https://benchgap.net/model/decision-index-0-3-tev1-4b-experimental.md) | 18.5% | 41.4% |
| [this-that 1.2](https://benchgap.net/model/decision-index-0-3-this-that-1-2.md) | 37.9% | 23.4% |
| [Torchcast Decision 27B](https://benchgap.net/model/decision-index-0-3-torchcast-decision-27b.md) | 63.2% | 57.3% |
| [Verdict](https://benchgap.net/model/decision-index-0-3-verdict.md) | 0.0% | 0.0% |
| [vLLM-SR Decision 2.0 Eos 0.8B](https://benchgap.net/model/decision-index-0-3-vllm-sr-decision-2-0-eos-0-8b.md) | 33.1% | 19.7% |
| [vLLM-SR Decision 2.0 Kai 0.6B](https://benchgap.net/model/decision-index-0-3-vllm-sr-decision-2-0-kai-0-6b.md) | 26.8% | 13.1% |
| [vLLM-SR Decision 2.0 Lux 9B](https://benchgap.net/model/decision-index-0-3-vllm-sr-decision-2-0-lux-9b.md) | 54.3% | 46.1% |
| [vLLM-SR Decision 2.0 Nox 4B](https://benchgap.net/model/decision-index-0-3-vllm-sr-decision-2-0-nox-4b.md) | 52.3% | 42.2% |
| [vLLM-SR Decision 2.0 Sol 2B](https://benchgap.net/model/decision-index-0-3-vllm-sr-decision-2-0-sol-2b.md) | 40.4% | 24.0% |
| [vLLM-SR Decision 2.0 Vega 27B](https://benchgap.net/model/decision-index-0-3-vllm-sr-decision-2-0-vega-27b.md) | 61.1% | 53.9% |
| [Von 1.3](https://benchgap.net/model/decision-index-0-3-von-1-3.md) | 20.9% | 14.7% |
| [Winnow-E4B](https://benchgap.net/model/decision-index-0-3-winnow-e4b.md) | 40.8% | 37.5% |
| [Xor 1.2](https://benchgap.net/model/decision-index-0-3-xor-1-2.md) | 57.9% | 53.9% |
| [Zev](https://benchgap.net/model/decision-index-0-3-zev.md) | 0.0% | 1.8% |
| [Decider chat · Gemma-4-31B](https://benchgap.net/model/decision-index-decider-chat-gemma4-31b.md) | 63.4% | 55.5% |
| [Surogate Rune 26B-A4B v3](https://benchgap.net/model/decision-index-rune-26b-a4b-v3.md) | 62.3% | 58.6% |
| [simple-jev · Qwen3.8-27B](https://benchgap.net/model/decision-index-simple-jev-qwen3-8-27b.md) | 63.5% | 56.1% |
| [djev](https://benchgap.net/model/djev.md) | 44.8% | 48.6% |
| [Drex Max](https://benchgap.net/model/drex-max.md) | 67.6% | 62.1% |
| [Drex v1.5](https://benchgap.net/model/drex-v1-5.md) | 59.2% | 40.8% |
| [Drex v2.0](https://benchgap.net/model/drex-v2-0.md) | 58.5% | 52.7% |
| [Eikos-27B](https://benchgap.net/model/eikos-27b.md) | 59.3% | 55.1% |
| [elephant-1](https://benchgap.net/model/elephant-1.md) | 65.1% | 62.1% |
| [GLiDE v1.2](https://benchgap.net/model/fastino-glide-v1-2.md) | 73.4% | 66.8% |
| [jeff](https://benchgap.net/model/jeff.md) | 17.2% | 7.8% |
| [Jev 1.14](https://benchgap.net/model/jev-1-14.md) | 64.8% | 58.0% |
| [Jev-Omni](https://benchgap.net/model/jev-omni.md) | 39.3% | 48.1% |
| [Laya](https://benchgap.net/model/laya.md) | 9.2% | 14.6% |
| [mammoth-1](https://benchgap.net/model/mammoth-1.md) | 66.2% | 67.1% |
| [metask-jev-4b](https://benchgap.net/model/metask-jev-4b.md) | 14.1% | 37.0% |
| [Perplexity Decider Flash v1.2 4B](https://benchgap.net/model/pplx-decider-flash-v1-2.md) | 62.2% | 62.2% |
| [Perplexity Decider Flash v1.3 4B](https://benchgap.net/model/pplx-decider-flash-v1-3.md) | 66.4% | 62.1% |
| [Perplexity Decider v1.2 27B](https://benchgap.net/model/pplx-decider-v1-2.md) | 69.4% | 65.5% |
| [Perplexity Decider v1.3 27B](https://benchgap.net/model/pplx-decider-v1-3.md) | 70.7% | 67.2% |
| [Winnow-12B Q8](https://benchgap.net/model/winnow-12b.md) | 54.6% | 50.6% |
