Jev Decision Index Leaderboard
decision175 models · Updated October 2026
On the Jev Decision Index benchmark, Fastino GLiDE ranks #1 with a score of 71.5. The full ranking, with cost per million tokens, is below.
| # | Model | Score | ECE | p50 latency | Input / 1M |
|---|---|---|---|---|---|
| 1 | Fastino GLiDEFastino | 71.5 | 0.020 | 93 ms | $0.15 |
| 2 | Lychee-27BLycheeAI | 67.3 | 0.108 | 97 ms | — |
| 3 | vLLM-SR d3 (27B)vLLM Semantic Router | 67.2 | 0.043 | 97 ms | — |
| 4 | linne-oak-1-0MISA-AICORE | 66.9 | 0.105 | 105 ms | — |
| 5 | Decider 31B (NVFP4)Mapika | 66.0 | 0.024 | 72 ms | — |
| 6 | bobcat-1.3sanghwa-na | 64.8 | 0.102 | 64 ms | — |
| 7 | amos-1-largeScira (mukaddamzaid) | 64.8 | 0.032 | 148 ms | — |
| 8 | frontier-infra Jebadiah 27B v2.1Frontier Infra | 64.4 | 0.056 | 303 ms | — |
| 9 | Sieve-31BSthānika AI | 64.3 | 0.033 | 253 ms | — |
| 10 | AnyJev L0 · Gemma-4-31Bmorriszjm | 63.6 | 0.240 | 98 ms | — |
| 11 | deck31b (Gemma 4 31B)krishna-gogineni-765 | 63.0 | 0.204 | 219 ms | — |
| 12 | Darwin-27B-ZTCFINAL_Bench | 63.0 | 0.045 | 104 ms | — |
| 13 | KnowLine-35B-A3B-Gen1PelaAI | 62.9 | 0.048 | 41 ms | — |
| 14 | Perplexity Decider v1.1 (27B)Perplexity | 62.8 | 0.071 | 104 ms | $0.02 |
| 15 | decisio 0.6.0 (Gemma 4 31B)aminry | 62.7 | 0.142 | 54 ms | — |
| 16 | Decider chat · Gemma-4-31BMapika | 61.6 | 0.047 | 109 ms | — |
| 17 | Torchcast Decision 27BTorchcast AI | 61.5 | 0.024 | 100 ms | — |
| 18 | Qwen Decision 27B noncausal LoRAjsaurabh | 61.1 | 0.051 | 127 ms | — |
| 19 | Quyet-1.0-Largechinhnc | 60.8 | 0.078 | 105 ms | — |
| 20 | JevTypeSafe AI | 60.2 | 0.080 | 196 ms | $0.04 |
| 21 | Sieve 27BSthānika AI | 60.2 | 0.065 | 101 ms | — |
| 22 | OneJev1.5-27BOmniJev | 59.9 | 0.024 | 224 ms | — |
| 23 | mogan-decision-31B-itMogan AI | 59.8 | 0.054 | 109 ms | — |
| 24 | Shisa DE-2 (repeat)Shisa.AI | 59.8 | 0.207 | 57 ms | — |
| 25 | linne-fern 1.0 (27B)MISA-AICORE | 59.6 | 0.079 | 108 ms | — |
| 26 | Kev 27BJared Palmer | 59.6 | 0.022 | 135 ms | — |
| 27 | Xor 26B-A4B v1.1Juspay Technologies | 59.4 | 0.068 | 110 ms | — |
| 28 | Matilda Jev v1.5Maincode | 59.2 | 0.077 | 89 ms | — |
| 29 | garas (Gemma 4 31B)plantwaterer | 59.0 | 0.049 | 117 ms | — |
| 30 | Eikos-27Bcaiovicentino1 | 58.8 | 0.057 | 130 ms | — |
| 31 | reflex Qwen3.8-27B-FP8 (pmscr)kshetrajna12 | 58.8 | 0.045 | 149 ms | — |
| 32 | RSI-Jev v6.1 VL 27B (effort auto)shgao | 58.6 | 0.077 | 76 ms | — |
| 33 | vLLM-SR Decision 2.0 Vega 27BvLLM Semantic Router | 58.0 | 0.086 | 120 ms | — |
| 34 | JEV-35BAutoTrust AI Lab | 57.6 | 0.085 | 51 ms | — |
| 35 | JEV-27BAutoTrust AI Lab | 57.4 | 0.076 | 102 ms | — |
| 36 | simple-jev · Qwen3.8-27B (featherless)featherless-ai | 57.0 | 0.113 | 373 ms | — |
| 37 | Surogate Rune 26B-A4B v3Surogate | 56.4 | 0.120 | 121 ms | — |
| 38 | SynACK Decide 26B-A4B v7SkyPanther | 56.2 | 0.069 | 129 ms | — |
| 39 | Blink v0.3 26B-A4B (NVFP4)Pixilab.ai | 55.9 | 0.040 | 25 ms | — |
| 40 | JADEtheunnecessarythings | 55.6 | 0.238 | 133 ms | — |
| 41 | Wald-26B-A4BORG2 AI | 55.4 | 0.043 | 90 ms | — |
| 42 | AJev lora1 (Gemma 4 26B-A4B)andyzhang232 | 55.3 | 0.048 | 89 ms | — |
| 43 | Cloudflare clefCloudflare | 54.9 | 0.040 | 102 ms | $0.04 |
| 44 | Decider chat · Qwen3.6-27BMapika | 54.7 | 0.021 | 84 ms | — |
| 45 | vLLM-SR d3-flash (9B)vLLM Semantic Router | 54.3 | 0.056 | 54 ms | — |
| 46 | JPT-35B-A3Bkirp | 54.1 | 0.049 | 250 ms | — |
| 47 | Xor 1.2Juspay Technologies | 53.6 | 0.024 | 150 ms | — |
| 48 | GEV-26B-DecideAutoTrust AI Lab | 53.5 | 0.072 | 129 ms | — |
| 49 | OneJev1.5-9BOmniJev | 53.3 | 0.029 | 111 ms | — |
| 50 | Nace.AI Drex v2.0 (35B-A3B)Nace.AI | 52.8 | 0.018 | 276 ms | $0.06 |
| 51 | Bespoke Nimble 9B v3Bespoke Labs | 51.1 | 0.123 | 74 ms | — |
| 52 | frontier-infra Jebadiah 9B v2.1Frontier Infra | 50.9 | 0.055 | 125 ms | — |
| 53 | decisio 0.4.0 (Qwen3.6-35B-A3B)aminry | 50.8 | 0.055 | 71 ms | — |
| 54 | vLLM-SR d3-mini (4B)vLLM Semantic Router | 50.0 | 0.056 | 50 ms | — |
| 55 | JoshuaSP diffusiongemma (open-jev)JoshuaSP | 49.5 | 0.215 | 260 ms | — |
| 56 | OneJev1.5-4BOmniJev | 49.2 | 0.034 | 113 ms | — |
| 57 | AJev lora7 (Gemma 4 12B)andyzhang232 | 48.9 | 0.082 | 69 ms | — |
| 58 | CoCo-Decision-4B 2.3.0Corners | 48.8 | 0.116 | 63 ms | — |
| 59 | decisio 0.4.0 (Gemma 4 12B)aminry | 48.8 | 0.103 | 38 ms | — |
| 60 | PurpleMIST-Flash-1.0Olaverse Lab | 47.5 | 0.032 | 53 ms | — |
| 61 | Sieve 9B v2Sthānika AI | 47.0 | 0.073 | 51 ms | — |
| 62 | Decider 35B-A3BMapika | 46.6 | 0.023 | 101 ms | — |
| 63 | Winnow-12BEldanRing | 46.6 | 0.168 | 73 ms | — |
| 64 | vLLM-SR Decision 2.0 Lux 9BvLLM Semantic Router | 46.0 | 0.056 | 37 ms | — |
| 65 | Nace.AI Drex v1.5Nace.AI | 45.8 | 0.090 | 54 ms | $0.04 |
| 66 | Hopper 12B (trained)HopitAI | 45.8 | 0.076 | 290 ms | — |
| 67 | Cloudflare clef-omni (30B-A3B)Cloudflare | 45.0 | 0.040 | 191 ms | $0.15 |
| 68 | JevK5 9B v0.3.3alibiserikbay | 44.9 | 0.060 | 40 ms | — |
| 69 | Cloudflare clef-flashCloudflare | 44.4 | 0.025 | 53 ms | $0.02 |
| 70 | Jevfirekikoncuo | 44.4 | 0.052 | 90 ms | — |
| 71 | RSI-Jev v6.1 VL 4Bshgao | 44.3 | 0.049 | 58 ms | — |
| 72 | jiwo 4B v2eljiwo | 44.1 | 0.068 | 50 ms | — |
| 73 | JPT-9Bkirp | 44.1 | 0.077 | 149 ms | — |
| 74 | ezjev 4B s2everettjf | 43.9 | 0.063 | 36 ms | — |
| 75 | Jev-Omniakhilaaa3 | 43.8 | 0.161 | 55 ms | — |
| 76 | Solomon v1.1Doccy Health | 43.2 | 0.081 | 224 ms | — |
| 77 | Standard One 8B SH (v3)Standard Thinking | 42.9 | 0.106 | 90 ms | — |
| 78 | JEV-9BAutoTrust AI Lab | 42.1 | 0.073 | 50 ms | — |
| 79 | Bespoke Nimble 9B v2Bespoke Labs | 42.1 | 0.024 | 77 ms | — |
| 80 | Sieve 9BSthānika AI | 41.7 | 0.043 | 50 ms | — |
| 81 | Sieve 4BSthānika AI | 41.0 | 0.031 | 93 ms | — |
| 82 | Kev 9B v2Jared Palmer | 40.9 | 0.045 | 60 ms | — |
| 83 | vLLM-SR Decision 2.0 Nox 4BvLLM Semantic Router | 40.8 | 0.030 | 34 ms | — |
| 84 | Aplomb 1EmpirioLabs AI | 40.8 | 0.075 | 54 ms | — |
| 85 | Hopper (G) 1.2HopitAI | 40.2 | 0.093 | 23 ms | — |
| 86 | FlyMy.AI Decision 4B v1.6FlyMy.AI | 39.8 | 0.046 | 27 ms | — |
| 87 | Standard One 3B SH (v2)Standard Thinking | 39.8 | 0.114 | 80 ms | — |
| 88 | JPT-4Bkirp | 39.6 | 0.079 | 139 ms | — |
| 89 | Intern-Decision-4BIntern Large Models | 38.7 | 0.028 | 44 ms | — |
| 90 | RSI-Jev v5.0 VL 3Bshgao | 38.6 | 0.048 | 19 ms | — |
| 91 | djevDavipar | 38.1 | 0.212 | 84 ms | — |
| 92 | razorback16 openjev diffusiongemma (NVFP4, vLLM)razorback16 | 37.6 | 0.232 | 38 ms | — |
| 93 | Valen-4BValen-Team | 37.6 | 0.107 | 167 ms | — |
| 94 | NeoHorse-Jev-4BTokenRhythm | 37.3 | 0.104 | 49 ms | — |
| 95 | JevK5 v0.3alibiserikbay | 36.6 | 0.060 | 26 ms | — |
| 96 | Decider 4BMapika | 36.3 | 0.084 | 13 ms | — |
| 97 | levInterfaze | 35.5 | 0.064 | 71 ms | — |
| 98 | Lev1 3.7Btheaviv | 35.4 | 0.058 | 29 ms | — |
| 99 | Kev 4B r10Jared Palmer | 34.9 | 0.032 | 59 ms | $0.04 |
| 100 | Saracura PT-BR 4BFelhen | 34.4 | 0.179 | 62 ms | — |
| 101 | Cam1cbuk10001 | 34.4 | 0.063 | 249 ms | — |
| 102 | vLLM-SR d3-nano (2B)vLLM Semantic Router | 34.3 | 0.112 | 38 ms | — |
| 103 | mmastrac diffusiongemma (vLLM PR 57250)mmastrac | 34.3 | 0.202 | 127 ms | — |
| 104 | Kas 4Bkpiya | 34.0 | 0.098 | 69 ms | — |
| 105 | Tev1-4B-experimentalTogether | 33.7 | 0.104 | 36 ms | $0.04 |
| 106 | Winnow-E4BEldanRing | 32.7 | 0.058 | 45 ms | — |
| 107 | Jevstral 8BAmirBraham | 32.2 | 0.047 | 28 ms | — |
| 108 | Metask-Jev-4BMetask-AI-Lab | 32.2 | 0.103 | 57 ms | — |
| 109 | Ryotide Qwen v0.3.1csabag | 32.0 | 0.092 | 81 ms | — |
| 110 | Intelif Qwen3-4BUserMoonlight | 31.3 | 0.071 | 23 ms | — |
| 111 | LiquidAI d1-3BLiquid AI | 31.1 | 0.047 | 16 ms | — |
| 112 | open-jev (pngwn)pngwn | 31.1 | 0.059 | 112 ms | — |
| 113 | Sieve 2B v2Sthānika AI | 30.8 | 0.095 | 68 ms | — |
| 114 | Jev-Style 2B v3chaoliangUNSW | 30.7 | 0.118 | 70 ms | — |
| 115 | JobeMantisShrimpdev | 30.4 | 0.123 | 54 ms | — |
| 116 | Gevva E4Bdavidburhans | 29.2 | 0.145 | 53 ms | — |
| 117 | OneJev1.5-0.8BOmniJev | 29.0 | 0.035 | 83 ms | — |
| 118 | openvonsgenai-craft | 28.8 | 0.371 | 21 ms | — |
| 119 | vera-corescar-ai | 28.1 | 0.122 | 100 ms | — |
| 120 | SemIfTheoLeeCJ | 26.7 | 0.099 | 113 ms | — |
| 121 | vLLM-SR d3-lite (0.8B)vLLM Semantic Router | 26.2 | 0.094 | 37 ms | — |
| 122 | jiwo 0.8B v2eljiwo | 25.8 | 0.068 | 36 ms | — |
| 123 | Saina Helm 2.0Saina | 25.3 | 0.168 | 39 ms | — |
| 124 | this-that 1.2FLock.io | 24.9 | 0.198 | 44 ms | — |
| 125 | Gevva E2Bdavidburhans | 24.3 | 0.090 | 43 ms | — |
| 126 | crivo-400mtardellirs | 23.7 | 0.121 | 17 ms | — |
| 127 | Decider 2BMapika | 23.6 | 0.077 | 8 ms | — |
| 128 | vLLM-SR Decision 2.0 Sol 2BvLLM Semantic Router | 23.6 | 0.037 | 26 ms | — |
| 129 | mini-jevr-ms | 22.8 | 0.346 | 67 ms | — |
| 130 | Valen-2BValen-Team | 22.2 | 0.105 | 128 ms | — |
| 131 | Sieve 2BSthānika AI | 21.9 | 0.088 | 63 ms | — |
| 132 | Intern-Decision-2BIntern Large Models | 20.9 | 0.061 | 34 ms | — |
| 133 | Sieve 0.8BSthānika AI | 20.8 | 0.088 | 52 ms | — |
| 134 | Strands Decider 2BStrands Agents | 20.7 | 0.044 | 52 ms | — |
| 135 | EXAONE-4.0-1.2B-JEV v0.3carrtesy | 20.6 | 0.079 | 149 ms | — |
| 136 | tasksource Decider nano v4 (150M)tasksource | 19.9 | 0.070 | 13 ms | — |
| 137 | vLLM-SR Decision 2.0 Eos 0.8BvLLM Semantic Router | 18.4 | 0.130 | 25 ms | — |
| 138 | Crivo-150Mtardellirs | 18.3 | 0.115 | 13 ms | — |
| 139 | Bekko System One v0 400Mhotchpotch | 18.0 | 0.111 | 17 ms | — |
| 140 | vLLM-SR d3-edge (0.6B)vLLM Semantic Router | 17.7 | 0.141 | 19 ms | — |
| 141 | DecisionTune 1.0DecisionTune | 17.2 | 0.147 | 14 ms | — |
| 142 | Valen-0.8BValen-Team | 16.6 | 0.108 | 128 ms | — |
| 143 | Bosun v3.1 1.7BHanno Labs | 16.5 | 0.130 | 40 ms | — |
| 144 | vLLM-SR Decision 2.0 Kai 0.6BvLLM Semantic Router | 16.4 | 0.038 | 19 ms | — |
| 145 | Kev 0.8B r15Jared Palmer | 16.3 | 0.040 | 45 ms | — |
| 146 | JPT-0.8Bkirp | 15.5 | 0.069 | 118 ms | — |
| 147 | jev-ettin-400m-20261013-v0Elfsong | 15.2 | 0.090 | 16 ms | — |
| 148 | Sifr 0.8B v3.1mohamedlotfy50 | 14.5 | 0.120 | 35 ms | — |
| 149 | Crivo-32Mtardellirs | 14.2 | 0.133 | 8 ms | — |
| 150 | Dinah-0Lukitaduarte | 13.9 | 0.104 | 14 ms | — |
| 151 | Kodiak v0.2 1B (accuracy)Cortex Agent LLC | 13.9 | 0.077 | 142 ms | — |
| 152 | Intern-Decision-0.8BIntern Large Models | 12.7 | 0.025 | 34 ms | — |
| 153 | Jev-Style 0.8B v3chaoliangUNSW | 12.5 | 0.084 | 36 ms | — |
| 154 | Tev1-0.8B-experimentalTogether | 12.5 | 0.126 | 29 ms | — |
| 155 | Bekko System One v0 68Mhotchpotch | 12.4 | 0.113 | 12 ms | — |
| 156 | LiquidAI d1-omni-600MLiquid AI | 11.8 | 0.128 | 11 ms | — |
| 157 | Vonwfzyx | 11.7 | 0.216 | 18 ms | — |
| 158 | Bosun v3.1 0.6BHanno Labs | 11.0 | 0.142 | 39 ms | — |
| 159 | vera-sparkscar-ai | 11.0 | 0.157 | 80 ms | — |
| 160 | LavoirMogan AI | 10.8 | 0.149 | 20 ms | — |
| 161 | GLiNER2.5-DecideFastino | 10.1 | 0.088 | 23 ms | — |
| 162 | LAYAConvai Innovations | 9.2 | 0.140 | 6 ms | $0.00 |
| 163 | JevK5 Litealibiserikbay | 8.8 | 0.149 | 19 ms | — |
| 164 | LFM2.5-2.6B-RLCDmonotykamary | 8.2 | 0.255 | 39 ms | — |
| 165 | Qwen-2.5-1B-RLCDharshatheg | 7.8 | 0.205 | 44 ms | — |
| 166 | Bekko System One v0 17Mhotchpotch | 7.7 | 0.161 | 7 ms | — |
| 167 | CLM-v0.1-8BContrastive-LM | 6.5 | 0.323 | 47 ms | — |
| 168 | jefflogan-markewich | 6.2 | 0.097 | 22 ms | — |
| 169 | LFM2.5-350M-RLCDnotnotsamuel | 5.9 | 0.568 | 27 ms | — |
| 170 | system-one-gemmaakash-kamat | 5.2 | 0.239 | 32 ms | — |
| 171 | DM-JEPA 1.1Danger Labs | 5.0 | 0.445 | 63 ms | — |
| 172 | Julia 1Supersonic Labs | 4.8 | 0.420 | 6 ms | — |
| 173 | Nirnaya 0.4Bkshitijthakkar | 3.5 | 0.046 | 14 ms | — |
| 174 | Zevbhubbard | 3.4 | 0.138 | 1 ms | — |
| 175 | RLCD ModernBERT 151Mheman10x | 2.0 | 0.154 | 11 ms | — |
Frequently asked questions
What is the Jev Decision Index benchmark?
Jev Decision Index is a decision used to evaluate and rank large language models. This page lists 175 models by their Jev Decision Index score.
Which model is best on Jev Decision Index?
Fastino GLiDE (Fastino) currently leads the Jev Decision Index leaderboard with a score of 71.5.
What is the best value model on Jev Decision Index?
Pricing data is being matched to these models.
Pricing is indicative — confirm with the provider before production use. Updated October 2026.