whatmodel.sh / decision / Nace.AI Drex v1.5

Nace.AI Drex v1.5

+ contribute

Nace.AI Drex v1.5 is an open-weights decision model from Nace.AI. Its published size is 8.95B parameters. Licence: other.

Catalogue identity notes

Exact Decision Index tested engine drex-v1.5; reported model Nace.AI Drex v1.5. Type: full fine-tune; technique: autoregressive. Comparison entry; no current run command, API price or speed inferred from the benchmark. Current repository metadata checked separately; an unpinned alias is not evidence that today’s weights reproduce this historical run.

Model information sources

· Weights licence ↗ · Parameter count ↗

Publisher parameter count; includes the served backbone, not just adapter size.

57.8±0.8#131 of 355 decision modelsfrom 86 benchmark results across 6 sources

$ how it compares

The decision models ranked either side of Nace.AI Drex v1.5, out of 357 with a score, then more from Nace.AI.

#	model	score	where	RAM	released
131	Winnow-12B (DecisionBench b6ac22b0) EldanRing	58±2	local	~16GB
132	deepseek-v4.1-flash (DecisionBench openrout) DeepSeek	58±2	model files
133	Nace.AI Drex v1.5 Nace.AI	58±1	model files
134	Clef-flash Cloudflare	58±1	api $0.038/M input tokens +local	~25GB	2026-10
135	Bespoke Nimble 9B v3 Bespoke Labs	58±1	local	~21GB	2026-10-06
	more from Nace.AI
78	Nace.AI Drex v2.0 (35B-A3B) Nace.AI	63±1	model files

see all decision models · find the ones that fit your machine

$ what it's good at

tools & automation▁▃▁▁▁▂▃▄▆██▆ #121 of 250 · 5.0% of its score · 8 tests

Beats Tern-35B-A3B, GEV-26B-Decide, Surogate Rune 26B-A4B v3 here, though they rank higher overall. Trails OneJev1.6-0.8B, TokenRhythm NeoHorse-Jev-4B, razorback16 diffgemma (decision readout) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
DecideBench v1: agent routing	98.0% ✎	▄·····▂▄·█·▄ beats 7/10	very easy · avg 93%	▇ 1.2%	sharp · maxed out · noisy
Decision Index 0.4.1: API-Bank (accuracy) submitted	85.6% ✎	▁▁▁▁▁▁▁▂▃▂█▂ beats 183/236	easy · avg 70%	▇ 0.9%	sharp · noisy
DecideBench v1: action review	90.0% ✎	▃▃·▃·█···█·▅ beats 5/10	easy · avg 87%	▇ 0.9%	sharp · noisy
Decision Index 0.4.1: BFCL (case exact accuracy) submitted	94.8% ✎	▁▂▁▁▁▂▂▄▅▆█▁ beats 126/246	very easy · avg 92%	▇ 0.7%	sharp · maxed out · noisy
Decision Index 0.4.1: Home appliance simulator (case exact accuracy) submitted	38.6% ✎	█▁▁▃▄▄▆▄▁▂▁▃ beats 113/243	medium · avg 37%	▇ 0.5%	sharp · at its level · noisy
Decision Index 0.4.1: When2Call (accuracy) submitted	89.6% ✎	▁▁▁▂▃▅▅▆█▇█▃ beats 212/246	easy · avg 72%	▇ 0.4%	noisy
Decision Index 0.4.1: ToolRet (nDCG@10) submitted	70.6% ✎	▁▁▁▁▁·▁▁▁▄█▃ beats 236/246	medium · avg 60%	▇ 0.4%	blunt · at its level
Decision Index 0.4.1: RouterBench (selected quality (quality objective)) submitted	79.9% ✎	no other model yet		—	failed run: Publisher excludes this protocol because its prompt reveals GPT-4 as the best route. Evidence: https://huggingface.co/spaces/multimodalart/jev-decision-index/blob/c130faf65f82e81d3bdefb01e758c712dc5c5199/data/index.json

retrieval & classification▁▁▁·▁▁▁▁▃▃█▅ #103 of 258 · 11% of its score · 16 tests

Beats AutoTrust-JEV-27B v1.1, Decider 31B (NVFP4), Quyet-1.0-Large here, though they rank higher overall. Trails Jev-Style 2B v4, Sieve 2B v3, razorback16 diffgemma (decision readout) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
System One Bench: MIRACL pairwise (Arabic)	86.0% ✎	▂···▁··▃▄▂█▇ beats 7/29	easy · avg 84%	▇ 1.7%	at its level
System One Bench: MASSIVE intent (Arabic)	77.5% ✎	▄·▁·▁▃▃▃█▄▇▇ beats 8/29	easy · avg 76%	▇ 1.6%	typical
LeanMCP · scifact rank noul	74.2% ✎	▂····▂···▃▅█ beats 7/11	medium · avg 62%	▇ 1.3%	at its level
Decision Index 0.4.1: BRIGHT (nDCG@10) submitted	44.5% ✎	▁▁▁▁▁▁▁▂▂▄▇█ beats 148/246	medium · avg 38%	▇ 1.1%	blunt · at its level
System One Bench: MIRACL rerank (Arabic)	89.0% ✎	▁▂·▁▁·▂·▄▄█▇ beats 25/29	easy · avg 79%	▇ 1.0%	typical
Decision Index 0.4.1: Amazon ESCI (macro-F1) submitted	56.2% ✎	▁▁▁▁▁▁▂▂▃▃█▅ beats 170/246	medium · avg 47%	▇ 1.0%	at its level
LeanMCP · nfcorpus rank noul	37.3% ✎	▃▃·····▃▃▅██ beats 7/11	hard · avg 29%	▇ 0.9%	blunt · at its level
System One Bench: MASSIVE scenario (Arabic)	58.5% ✎	▃▁·▄▃▂█▂▁▁·▃ beats 8/29	medium · avg 60%	▇ 0.6%	at its level

8 more tests

test	result	vs the others in its run	difficulty	counts	why
DecideBench v1: support intent	98.0% ✎	▃····▂▂·▃··█ beats 4/10	very easy · avg 95%	▇ 0.4%	sharp · maxed out · noisy
Decision Index 0.4.1: HoVer (accuracy) submitted	86.1% ✎	▆▆▇▅▄▄▃▇▄█▆▄ beats 194/246	easy · avg 74%	▇ 0.3%	blunt
Decision Index 0.4.1: PhishNChips phishing decisions (accuracy) submitted	66.5% ✎	█▅▅▄▃▂▁▁▁▁·▁ beats 111/245	easy · avg 70%	▇ 0.2%	noisy
System One Bench: AG News (English)	90.0% ✎	▁···▅▁▅▂▆█▂▅ beats 18/29	easy · avg 88%	▇ 0.2%	blunt
Decision Index 0.4.1: CLINC150+OOS (macro-F1) submitted	91.8% ✎	▁▁▁▁▁▁▁▂▃▆█▄ beats 152/238	easy · avg 83%	▇ 0.2%	sharp · noisy
Decision Index 0.4.1: BANKING77 (macro-F1) submitted	85.4% ✎	▁·▁▁▁▁▁▁▂█▇█ beats 134/238	easy · avg 79%	▇ 0.1%	noisy
System One Bench: Banking77 (English)	87.0% ✎	█▅▆█▅▃▃▃···▂ beats 11/26	very easy · avg 87%	▇ 0.0%	blunt · noisy
Decision Index 0.4.1: SGD/SGD-X (macro-F1) submitted	65.5% ✎	no other model yet		—	failed run: Publisher reports a builder bug exposing dialogue turns after the classified turn; awaiting corrected rerun. Evidence: https://huggingface.co/spaces/multimodalart/jev-decision-index/blob/c130faf65f82e81d3bdefb01e758c712dc5c5199/data/index.json

language understanding▂▂▂▃▂▅▆▃█▆▄▃ #112 of 256 · 18% of its score · 21 tests

Beats jev-brain · Gemma-4-31B (decision readout), Decider chat · Gemma-4-31B (decision readout), garum (Gemma 4 31B) here, though they rank higher overall. Trails Decider 4B, jiwo 4B v2, Sieve 4B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: ANLI (macro-F1) submitted	70.3% ✎	▁▂▃▅▅█▂····▁ beats 162/246	medium · avg 60%	▇ 2.3%	at its level
Decision Index 0.4.1: NLI4CT (macro-F1) submitted	80.3% ✎	▁▁▁▂▂▃▃▄▇▇█▄ beats 154/246	easy · avg 74%	▇ 2.2%	typical
DecideBench v1: review sentiment	90.0% ✎	▂··▂···▂▃█▄▂ beats 3/10	easy · avg 85%	▇ 2.0%	sharp
Decision Index 0.4.1: HellaSwag (accuracy) submitted	92.1% ✎	▂▄▂▂▄▇▃▇█▅▃▁ beats 142/246	very easy · avg 85%	▇ 1.9%	sharp
DecideBench v1: content moderation	80.0% ✎	▂··▂···▄▄▂·█ beats 2/10	easy · avg 83%	▇ 1.3%	sharp · noisy
System One Bench: BoolQ (English)	89.0% ✎	▁▁▁▃·▁▃▃▇▇█▅ beats 13/29	easy · avg 86%	▇ 1.2%	typical
System One Bench: OSACT4 offensive (Arabic)	90.5% ✎	▁······▁▃█▇▄ beats 17/29	easy · avg 84%	▇ 1.1%	sharp
Decision Index 0.4.1: iSarcasmEval (Sarcasm F1 · track A, English) submitted	55.2% ✎	▁·▁·▁▁▂▄▅█▇▄ beats 161/246	medium · avg 44%	▇ 1.0%	at its level

13 more tests

test	result	vs the others in its run	difficulty	counts	why
System One Bench: SST-2 (English)	94.5% ✎	▁·▃▁▂▂▃▅█·▂▁ beats 13/29	easy · avg 92%	▇ 0.7%	typical
System One Bench: SST-5 (English)	53.5% ✎	▃·▁▁▃·▇▄▄▄█▅ beats 12/29	medium · avg 51%	▇ 0.6%	blunt · at its level
DecideBench v1: claim support	86.0% ✎	▄··▄·▄·▄▄███ beats 3/10	easy · avg 82%	▇ 0.6%	sharp · noisy
Decision Index 0.4.1: ACOS (per-review F1) submitted	5.0% ✎	▁··▂▃▃▅▇█▇▃▃ beats 25/246	hard · avg 22%	▇ 0.5%	at its level · noisy
Decision Index 0.4.1: ContractNLI (macro-F1) submitted	83.2% ✎	▁▁▁▁▁▁▁▂▂▇▇█ beats 199/246	easy · avg 71%	▇ 0.4%	noisy
Decision Index 0.4.1: VAST (macro-F1) submitted	77.0% ✎	▁▁▁▂▅▄▄▂▄▆█▄ beats 183/246	medium · avg 62%	▇ 0.4%	at its level
Decision Index 0.4.1: RAGTruth (F1 on hallucinated class) submitted	82.9% ✎	▁▁▁▁▁▁▁▁▁▃▆█ beats 210/246	easy · avg 66%	▇ 0.3%	at its level · noisy
System One Bench: XNLI Arabic (Arabic)	74.0% ✎	▄▁▃▁▄▇█▅▃▁·▃ beats 17/29	medium · avg 68%	▇ 0.3%	at its level
System One Bench: XNLI French (French)	82.0% ✎	▂▅·▁▁▅▅█▃▁·▃ beats 17/29	medium · avg 76%	▇ 0.3%	blunt · at its level
Decision Index 0.4.1: FinEntity (macro-F1) submitted	88.0% ✎	▁···▁▁▁▁▃▆█▃ beats 125/246	easy · avg 84%	▇ 0.3%	noisy
System One Bench: Tweet sentiment (Arabic)	63.5% ✎	▁······▁▁▄█▄ beats 9/29	medium · avg 62%	▇ 0.2%	blunt · at its level
System One Bench: XNLI English (English)	88.7% ✎	▃·▁·▃▇█▄█▃▃▁ beats 16/29	easy · avg 85%	▇ 0.1%	blunt · noisy
System One Bench: DAIR Emotion (English)	53.5% ✎	██▃·····▁··▁ beats 1/29	medium · avg 59%	▇ 0.0%	blunt · at its level

arts & human taste▁▇▄▃▅▅█▅▄▂▁▁ #151 of 247 · 4.0% of its score · 6 tests

Trails Crivo-32M v2, Crivo-68M, Crivo-150M v2 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: New Yorker (accuracy) submitted	72.9% ✎	▁▁▂▃▃▃▇▅█▅▃▁ beats 171/246	medium · avg 61%	▇ 1.7%	at its level
Decision Index 0.4.1: Humicroedit (accuracy) submitted	63.0% ✎	▁·▁▃▂▂▅█▄▂▁▁ beats 174/246	medium · avg 60%	▇ 0.8%	blunt · at its level
Decision Index 0.4.1: cfcolor (accuracy) submitted	58.3% ✎	▁········▁▇█ beats 83/246	medium · avg 60%	▇ 0.5%	blunt · at its level
Decision Index 0.4.1: BPoMP (accuracy) submitted	82.2% ✎	▁█▄▄▅█▆▅▄▁▁▁ beats 96/246	easy · avg 83%	▇ 0.5%	sharp · noisy
Decision Index 0.4.1: POP909-CL (accuracy) submitted	76.3% ✎	▄▅▅▃▆▅▃▅▂▅▃█ beats 201/237	hard · avg 21%	▇ 0.4%	sharp · at its level · noisy
Decision Index 0.4.1: Habermas Machine (accuracy) submitted	59.7% ✎	▁▂▄█▇▅▁▁▄▄▁▁ beats 199/246	medium · avg 47%	▇ 0.1%	blunt · at its level

knowledge & reasoning▁▂▂▂▂▃▃▅▅▆█▂ #102 of 263 · 40% of its score · 21 tests

Beats Senova-4B v2, Verdin-4B v1.1, Cerno 9B v1.1 here, though they rank higher overall. Trails Decider 35B-A3B, decisio 0.4.0 (Qwen3.6-35B-A3B) (decision readout), Solar Decide here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Etsabary · 1342 text decisions	63.6% ✎	▁··▂·▃▅▁▃▂▂█ beats 3/24	medium · avg 47%	▇▇▇ 7.9%	sharp · at its level
Decision Index 0.4.1: MMLU-Pro (accuracy) submitted	58.7% ✎	▄▄▃▃▂▄▅▇██▅▁ beats 147/246	medium · avg 46%	▇▇ 4.5%	sharp · at its level
LeanMCP · scienceqa text	90.6% ✎	▄··▄█▄█▄·█▄▄ beats 1/11	very easy · avg 88%	▇ 3.2%	sharp · maxed out
Decision Index 0.4.1: MMLU (accuracy) submitted	78.5% ✎	▁▄▄▃▃▂▅▇▅█▇▁ beats 143/246	easy · avg 70%	▇ 2.9%	sharp
LeanMCP · medqa usmle	75.5% ✎	▄▄▄·▄▄▄▄▄·██ beats 1/11	easy · avg 72%	▇ 2.7%	typical
Decision Index 0.4.1: BBH (accuracy) submitted	85.6% ✎	▄▆▄▄█▇▇▄▃▁▁▁ beats 217/246	easy · avg 66%	▇ 2.7%	at its level
MedQA · Freyer et al. (first-valid accuracy)	73.5% ✎	█··▃▃·▃▃▃·▃█ beats 1/11	easy · avg 72%	▇ 2.1%	typical
LeanMCP · medmcqa	66.6% ✎	▃·▃··▃█▃·▅▅▃ beats 1/11	medium · avg 63%	▇ 1.8%	at its level

13 more tests

test	result	vs the others in its run	difficulty	counts	why
LeanMCP · pubmedqa	65.2% ✎	▂···▂·▂·▄▄▂█ beats 3/11	medium · avg 54%	▇ 1.8%	at its level
Decision Index 0.4.1: GSM8K (accuracy (4- and 10-choice)) submitted	54.6% ✎	▁··▄█▆▆▆▅▅▃▁ beats 148/246	medium · avg 45%	▇ 1.5%	sharp · at its level
Decision Index 0.4.1: GPQA Diamond (accuracy) submitted	44.4% ✎	▂▆▅▅▆█▆▃▁··▁ beats 149/246	medium · avg 39%	▇ 1.4%	blunt · at its level
Decision Index 0.4.1: ARC-Challenge (accuracy) submitted	95.8% ✎	▂▂▂▂▂▂▁▃▄▆█▆ beats 157/246	very easy · avg 90%	▇ 1.4%	sharp · maxed out
LeanMCP · mmlu pro	61.0% ✎	▄··▄▄████··▄ beats 2/11	medium · avg 54%	▇ 1.3%	at its level
Decision Index 0.4.1: WinoGrande (accuracy) submitted	87.8% ✎	█▅█▅▄█▆▄▃▃▁▁ beats 179/246	easy · avg 78%	▇ 1.3%	typical
Decision Index 0.4.1: ARC-Easy (accuracy) submitted	97.9% ✎	▁▁▂▁▂▂▂▂▃▆█▅ beats 123/246	very easy · avg 96%	▇ 0.7%	sharp · maxed out · noisy
Decision Index 0.4.1: SATA-Bench (case exact accuracy) submitted	26.7% ✎	▁▁▁▁▁▁▁▂▅█▂▁ beats 115/246	hard · avg 22%	▇ 0.6%	at its level
Decision Index 0.4.1: MuSR (accuracy) submitted	57.9% ✎	▁▁▂█▅▄▁▁▁▁▁▁ beats 103/246	medium · avg 60%	▇ 0.5%	blunt · at its level
Decision Index 0.4.1: CRUXEval (accuracy) submitted	76.0% ✎	██▄▃▃▄▁▁···▁ beats 185/246	medium · avg 58%	▇ 0.5%	at its level · noisy
Decision Index 0.4.1: CLadder (accuracy) submitted	86.3% ✎	▁███▆▅▃▃▁▁▁▁ beats 208/246	easy · avg 71%	▇ 0.4%	typical
Decision Index 0.4.1: ChessBench (accuracy) submitted	21.9% ✎	▄▇▄▄▄▆▄█▄▁▁▁ beats 186/246	hard · avg 16%	▇ 0.3%	blunt
Decision Index 0.4.1: HLE (accuracy) submitted	11.2% ✎	▁▁▂▃▆█▇▅▅▃▁▁ beats 72/246	very hard · avg 13%	▇ 0.0%	blunt

workflows▁▁▁▂▁▂▂▆█▄▁▁ #41 of 114 · 19% of its score · 14 tests

Beats Clef-omni, Jebadiah 27B v2.1 (Ollaya Q4_K_M), vLLM-SR d3-mini (4B) here, though they rank higher overall. Trails StartLux-Decision v1.1 2B, Solar Decide, jqv (DecisionBench 9216db57) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
LeanMCP · mcpbench server routing	92.9% ✎	█·███··▄···▄ beats 3/9	easy · avg 82%	▇ 3.4%	sharp
LeanMCP · mcpbench tool selection	100.0% ✎	▂··········█ beats 3/9	very easy · avg 89%	▇ 2.2%	sharp · maxed out · noisy
DecideBench v1: ticket triage	94.0% ✎	▄·····▂█·▆·▂ beats 5/10	very easy · avg 90%	▇ 1.8%	sharp · maxed out
LeanMCP · rewardbench2	69.3% ✎	▄▄▄·▄▄▄▄█·█▄ beats 4/11	medium · avg 58%	▇ 1.6%	at its level
LeanMCP · mcpmark server routing	32.8% ✎	▃··▂··▄█▂·▂▄ beats 0/11	medium · avg 36%	▇ 1.5%	at its level
System One Bench: Typed decisions (English)	64.8% ✎	▁▁·▁▁▁▅▄▇██▁ beats 7/29	medium · avg 63%	▇ 1.4%	blunt · at its level
LeanMCP · aegis2 response	72.8% ✎	█··▃·▃·█··▅▅ beats 2/11	easy · avg 70%	▇ 1.3%	typical
LeanMCP · mcpmark tool selection	100.0% ✎	▂··········█ beats 3/11	very easy · avg 96%	▇ 1.3%	sharp · maxed out · noisy

6 more tests

test	result	vs the others in its run	difficulty	counts	why
LeanMCP · atbench500	77.0% ✎	▃··▅▅█··▅··▅ beats 7/11	medium · avg 63%	▇ 1.0%	sharp · at its level · noisy
LeanMCP · helpsteer2 helpfulness	1.0483 ordinal levels ✎	▄▄▄▄▄·▄·█▄▄▄ beats 4/10	avg 1.357 ordinal levels	▇ 0.9%	blunt
DecideBench v1: returns policy	84.0% ✎	▃··▃··▃·▅▅█▃ beats 3/10	easy · avg 81%	▇ 0.9%	sharp · noisy
LeanMCP · sst5	0.5590 ordinal levels ✎	▂·····▂▃▄█▄▂ beats 2/11	avg 0.879 ordinal levels	▇ 0.9%	blunt
LeanMCP · prompt injections	55.2% ✎	▃▂▂▂█·▄▂···▂ beats 0/11	medium · avg 62%	▇ 0.7%	at its level
LeanMCP · jailbreak classification	92.0% ✎	▂·▄··▂···█▆▂ beats 1/11	very easy · avg 95%	▇ 0.0%	blunt · maxed out · noisy

other tests2 not in a task · 3.7% of its score

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: private same skills (equated points)	49.1700 points ✎	▂▂▂▃▃▃▄▅▅▇█▃ beats 140/247	avg 41.206 points	▇ 2.4%	blunt
Decision Index 0.4.1: private new domains (equated points)	40.7000 points ✎	▃█▃▃▂▆▃▃▄▄▄▄ beats 132/247	avg 38.861 points	▇ 1.3%	blunt

$ where to run it

Download model files ↗
Published model files verified. This is a verified repository snapshot; the benchmark does not establish an identical weight revision. A local run command and hardware requirements have not been verified for this entry. Source · checked 2026-10-31.

Model repository and documentation: huggingface.co/nace-ai ↗

$ how the score was worked out · Decision Index 0.4.1 42%, LeanMCP · JevBench 28%, System One Bench (Ali Alhashi… 11% and 3 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
Decision Index 0.4.1 · submitted; independent execution unverified	42	▇▇▇▇▇▇▇▇▇▇▇▇▇ 42%	CC BY 4.0
LeanMCP · JevBench	18	▇▇▇▇▇▇▇▇ 28%	MIT
System One Bench (Ali Alhashimi)	16	▇▇▇ 11%	CC-BY-4.0
DecideBench (Cho Yin Yong)	8	▇▇▇ 9%	CC BY 4.0
Etsabary · decision benchmarks	1	▇▇ 8%	MIT
Freyer et al. · MedQA	1	▇ 2%	CC BY 4.0
prior · pull toward the middle		▇ 0%
weighted average		= 57.8

not counted (2)

Decision Index 0.4.1	Decision Index 0.4.1: RouterBench (selected quality (quality objective)) [correct]	79.9%	failed run: Publisher excludes this protocol because its prompt reveals GPT-4 as the best route. Evidence: https://huggingface.co/spaces/multimodalart/jev-decision-index/blob/c130faf65f82e81d3bdefb01e758c712dc5c5199/data/index.json
Decision Index 0.4.1	Decision Index 0.4.1: SGD/SGD-X (macro-F1) [correct]	65.5%	failed run: Publisher reports a builder bug exposing dialogue turns after the classified turn; awaiting corrected rerun. Evidence: https://huggingface.co/spaces/multimodalart/jev-decision-index/blob/c130faf65f82e81d3bdefb01e758c712dc5c5199/data/index.json

$ improve this page

Update model information Add a place to run it Add a test result Propose a benchmark Correct a benchmark score Edit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated 2026-11-03 02:37 UTC · sources and licences · json