whatmodel.sh / decision / LiquidAI d1-3B

LiquidAI d1-3B

+ contribute

LiquidAI d1-3B is an open-weights decision model from Liquid AI. Its published size is 3.12B parameters. Licence: other.

Catalogue identity notes

Exact Decision Index tested engine d1-3b; reported model LiquidAI d1-3B. 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.

50.9±1.0#196 of 355 decision modelsfrom 68 benchmark results across 4 sources

$ how it compares

The decision models ranked either side of LiquidAI d1-3B, out of 357 with a score, then more from Liquid AI.

#	model	score	where	RAM	released
196	ZTC-Judge-9B (DecisionBench 05d154a5) FINAL-Bench	51±2	model files
197	Jev-Omni akhilaaa3	51±1	model files		2026-09-22
198	LiquidAI d1-3B Liquid AI	51±1	model files
199	djev-spark (DecisionBench 1444f3e9) mmastrac	51±2	source code
200	vLLM-SR Decision 2.0 Nox 4B vLLM Semantic Router	51±1	local	~12GB
	more from Liquid AI
21	d1-ultra Liquid AI	69±2	api price unknown		2026-10-15
32	d1.1 (hosted) Liquid AI	68±2	api price unknown
61	d1 (hosted) Liquid AI	65±1	api price unknown

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

$ what it's good at

tools & automation▁▃▁▁▁▂▃▄▆██▆ #179 of 250 · 5.5% of its score · 6 tests

Trails JevK5 Lite, Mola-150M, Gevva E2B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: API-Bank (accuracy) submitted	78.1% ✎	▁▁▁▁▁▁▁▂▃▃█▂ beats 110/236	easy · avg 70%	▇ 1.7%	sharp · at its level · noisy
Decision Index 0.4.1: BFCL (case exact accuracy) submitted	89.4% ✎	▁▂▁▁▁▂▂▄▅▆█▁ beats 70/246	very easy · avg 92%	▇ 1.7%	sharp · maxed out · noisy
Decision Index 0.4.1: When2Call (accuracy) submitted	86.6% ✎	▁▁▁▂▃▅▅▆█▇█▃ beats 186/246	easy · avg 72%	▇ 0.8%	at its level · noisy
Decision Index 0.4.1: Home appliance simulator (case exact accuracy) submitted	63.6% ✎	█▁▁▃▄▄▆▄▁▂▁▃ beats 155/243	medium · avg 37%	▇ 0.7%	sharp · at its level · noisy
Decision Index 0.4.1: ToolRet (nDCG@10) submitted	65.7% ✎	▁▁▁▁▁·▁▁▁▄█▁ beats 156/246	medium · avg 60%	▇ 0.6%	blunt · at its level
Decision Index 0.4.1: RouterBench (selected quality (quality objective)) submitted	79.5% ✎	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▁▁▁·▁▁▁▁▃▃█▅ #143 of 258 · 13% of its score · 13 tests

Beats Quyet-1.0-Large, Decider chat · Qwen3.6-27B (decision readout), Sieve 9B v2 here, though they rank higher overall. Trails Amber 0.4B, vLLM-SR vela3-fast (Vela 3.0, 0.4B), crystal-0.8b here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
System One Bench: MIRACL pairwise (Arabic)	89.0% ✎	▂···▁··▃▄▂█▇ beats 13/29	easy · avg 84%	▇ 2.8%	at its level
System One Bench: MASSIVE intent (Arabic)	83.5% ✎	▄·▁·▁▃▃▃█▄▇▇ beats 11/29	easy · avg 76%	▇ 2.8%	typical
System One Bench: MIRACL rerank (Arabic)	82.5% ✎	▁▂·▁▁·▂·▄▄█▇ beats 8/29	easy · avg 79%	▇ 1.7%	typical
Decision Index 0.4.1: BRIGHT (nDCG@10) submitted	40.6% ✎	▁▁▁▁▁▁▁▂▂▄▇█ beats 102/246	medium · avg 38%	▇ 1.5%	blunt · at its level
Decision Index 0.4.1: Amazon ESCI (macro-F1) submitted	51.2% ✎	▁▁▁▁▁▁▂▂▃▃█▅ beats 112/246	medium · avg 47%	▇ 1.4%	at its level
System One Bench: MASSIVE scenario (Arabic)	71.5% ✎	▃▁·▄▃▂█▂▁▁·▃ beats 17/29	medium · avg 60%	▇ 1.0%	at its level
Decision Index 0.4.1: HoVer (accuracy) submitted	74.1% ✎	▆▆▇▅▄▄▃▇▄█▆▄ beats 118/246	easy · avg 74%	▇ 0.4%	blunt · at its level
Decision Index 0.4.1: PhishNChips phishing decisions (accuracy) submitted	53.8% ✎	█▅▅▄▃▂▁▁▁▁·▁ beats 64/245	easy · avg 70%	▇ 0.4%	at its level · noisy

5 more tests

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: CLINC150+OOS (macro-F1) submitted	91.2% ✎	▁▁▁▁▁▁▁▂▃▆█▄ beats 145/238	easy · avg 83%	▇ 0.3%	sharp · noisy
System One Bench: AG News (English)	88.5% ✎	▁···▅▁▅▃▆█▂▅ beats 10/29	easy · avg 88%	▇ 0.3%	blunt
Decision Index 0.4.1: BANKING77 (macro-F1) submitted	88.0% ✎	▁·▁▁▁▁▁▁▂█▇█ beats 161/238	easy · avg 79%	▇ 0.2%	noisy
System One Bench: Banking77 (English)	88.0% ✎	█▅▆█▅▃▃▃···▂ beats 12/26	very easy · avg 87%	▇ 0.0%	blunt · noisy
Decision Index 0.4.1: SGD/SGD-X (macro-F1) submitted	69.2% ✎	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▂▂▂▃▂▅▆▃█▆▄▃ #128 of 256 · 23% of its score · 18 tests

Beats Bespoke Nimble 9B v3, Jebadiah 9B v2.1 (T1e), Solar Decide here, though they rank higher overall. Trails vLLM-SR d3-nano (2B), JevK5 v0.3, Decider 4B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: HellaSwag (accuracy) submitted	89.2% ✎	▂▄▂▂▄▇▃▇█▅▃▁ beats 127/246	very easy · avg 85%	▇▇ 4.3%	sharp
Decision Index 0.4.1: NLI4CT (macro-F1) submitted	74.1% ✎	▁▁▁▂▂▃▃▄▇▇█▄ beats 94/246	easy · avg 74%	▇▇ 3.8%	at its level
Decision Index 0.4.1: ANLI (macro-F1) submitted	62.4% ✎	▁▂▃▅▅█▂····▁ beats 112/246	medium · avg 60%	▇ 3.6%	at its level
System One Bench: OSACT4 offensive (Arabic)	84.5% ✎	▁······▁▃█▇▄ beats 7/29	easy · avg 84%	▇ 2.0%	sharp · at its level
System One Bench: BoolQ (English)	85.0% ✎	▁▁▁▃·▃▃▃▇▇█▅ beats 5/29	easy · avg 86%	▇ 1.9%	at its level
Decision Index 0.4.1: iSarcasmEval (Sarcasm F1 · track A, English) submitted	36.0% ✎	▁·▁·▁▁▂▄▅█▇▄ beats 76/246	medium · avg 44%	▇ 1.5%	at its level
System One Bench: SST-2 (English)	95.5% ✎	▁·▃▁▂▂▃▅█·▂▁ beats 23/29	easy · avg 92%	▇ 1.3%	typical
System One Bench: SST-5 (English)	56.0% ✎	▃·▁▁▃·▇▄▄▄█▅ beats 15/29	medium · avg 51%	▇ 0.8%	blunt · at its level

10 more tests

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: ContractNLI (macro-F1) submitted	80.6% ✎	▁▁▁▁▁▁▁▂▂▇▇█ beats 169/246	easy · avg 71%	▇ 0.7%	at its level · noisy
Decision Index 0.4.1: VAST (macro-F1) submitted	77.0% ✎	▁▁▁▂▅▄▄▂▄▆█▄ beats 182/246	medium · avg 62%	▇ 0.7%	at its level
Decision Index 0.4.1: ACOS (per-review F1) submitted	37.7% ✎	▁··▂▃▃▅▇█▇▃▃ beats 192/246	hard · avg 22%	▇ 0.7%	noisy
Decision Index 0.4.1: RAGTruth (F1 on hallucinated class) submitted	77.6% ✎	▁▁▁▁▁▁▁▁▁▃▆█ beats 152/246	easy · avg 66%	▇ 0.5%	at its level · noisy
Decision Index 0.4.1: FinEntity (macro-F1) submitted	88.0% ✎	▁···▁▁▁▁▃▆█▃ beats 124/246	easy · avg 84%	▇ 0.5%	noisy
System One Bench: XNLI Arabic (Arabic)	73.3% ✎	▄▁▃▁▄▇█▅▃▁·▃ beats 15/29	medium · avg 68%	▇ 0.4%	at its level
System One Bench: XNLI French (French)	78.7% ✎	▂▅·▁▁▅▅█▃▁·▃ beats 11/29	medium · avg 76%	▇ 0.4%	blunt · at its level
System One Bench: Tweet sentiment (Arabic)	63.5% ✎	▁······▁▁▄█▄ beats 8/29	medium · avg 62%	▇ 0.3%	blunt · at its level
System One Bench: XNLI English (English)	86.0% ✎	▃·▁·▃▇█▄█▃▃▁ beats 10/29	easy · avg 85%	▇ 0.1%	blunt · noisy
System One Bench: DAIR Emotion (English)	63.5% ✎	██▃·····▁··▁ beats 26/29	medium · avg 59%	▇ 0.0%	blunt · at its level

arts & human taste▁▇▄▃▅▅█▅▄▂▁▁ #150 of 247 · 6.2% of its score · 6 tests

Beats Nace.AI Drex v1.5, Jegue 0.1 (9B), Standard One 8B SH (v4) here, though they rank higher overall. 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	64.0% ✎	▁▁▂▃▃▃▇▅█▅▃▁ beats 123/246	medium · avg 61%	▇ 2.7%	at its level
Decision Index 0.4.1: Humicroedit (accuracy) submitted	63.6% ✎	▁·▁▃▂▂▅█▄▂▁▁ beats 185/246	medium · avg 60%	▇ 1.2%	blunt · at its level
Decision Index 0.4.1: BPoMP (accuracy) submitted	82.4% ✎	▁█▄▄▅█▆▅▄▁▁▁ beats 97/246	easy · avg 83%	▇ 0.9%	sharp · noisy
Decision Index 0.4.1: cfcolor (accuracy) submitted	62.0% ✎	▁········▁▇█ beats 125/246	medium · avg 60%	▇ 0.7%	blunt · at its level
Decision Index 0.4.1: POP909-CL (accuracy) submitted	25.3% ✎	▄▅▅▃▆▅▃▅▂▅▃█ beats 137/237	hard · avg 21%	▇ 0.5%	sharp · noisy
Decision Index 0.4.1: Habermas Machine (accuracy) submitted	58.8% ✎	▁▂▄█▇▅▁▁▄▄▁▁ beats 193/246	medium · avg 47%	▇ 0.2%	blunt · at its level

knowledge & reasoning▁▂▂▂▂▃▃▅▅▆█▂ #176 of 263 · 31% of its score · 14 tests

Beats baselabs joey-1 (2B) here, though they rank higher overall. Trails openvons Qwen3-4B (decision readout), Jevstral 8B, Jobe Qwen3.5-4B (decision readout) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: MMLU-Pro (accuracy) submitted	44.9% ✎	▄▄▃▃▂▄▅▇██▅▁ beats 97/246	medium · avg 46%	▇▇▇ 6.7%	sharp · at its level
Decision Index 0.4.1: MMLU (accuracy) submitted	66.6% ✎	▁▄▄▃▃▂▅▇▅█▇▁ beats 82/246	easy · avg 70%	▇▇ 5.1%	sharp · at its level
Decision Index 0.4.1: BBH (accuracy) submitted	72.4% ✎	▄▆▄▄█▇▇▄▃▁▁▁ beats 147/246	easy · avg 66%	▇▇ 4.4%	at its level
Decision Index 0.4.1: ARC-Challenge (accuracy) submitted	88.0% ✎	▂▂▂▂▂▂▁▃▄▆█▆ beats 82/246	very easy · avg 90%	▇ 3.5%	sharp
Decision Index 0.4.1: WinoGrande (accuracy) submitted	77.7% ✎	█▅█▅▄█▆▄▃▃▁▁ beats 118/246	easy · avg 78%	▇ 2.3%	typical
Decision Index 0.4.1: GSM8K (accuracy (4- and 10-choice)) submitted	32.7% ✎	▁··▄█▆▆▆▅▅▃▁ beats 83/246	medium · avg 45%	▇ 2.2%	sharp · at its level
Decision Index 0.4.1: ARC-Easy (accuracy) submitted	95.4% ✎	▁▁▂▁▂▂▂▃▃▆█▅ beats 84/246	very easy · avg 96%	▇ 2.0%	sharp · maxed out · noisy
Decision Index 0.4.1: GPQA Diamond (accuracy) submitted	28.6% ✎	▂▆▅▅▆█▆▃▁··▁ beats 51/246	medium · avg 39%	▇ 2.0%	blunt · at its level

6 more tests

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: SATA-Bench (case exact accuracy) submitted	31.3% ✎	▁▁▁▁▁▁▁▂▅█▂▁ beats 163/246	hard · avg 22%	▇ 0.8%	typical
Decision Index 0.4.1: MuSR (accuracy) submitted	56.8% ✎	▁▁▂█▅▄▁▁▁▁▁▁ beats 95/246	medium · avg 60%	▇ 0.8%	blunt · at its level
Decision Index 0.4.1: CRUXEval (accuracy) submitted	31.1% ✎	██▄▃▃▄▁▁···▁ beats 6/246	medium · avg 58%	▇ 0.7%	at its level · noisy
Decision Index 0.4.1: CLadder (accuracy) submitted	56.6% ✎	▁███▆▅▃▃▁▁▁▁ beats 53/246	easy · avg 71%	▇ 0.7%	at its level
Decision Index 0.4.1: ChessBench (accuracy) submitted	15.8% ✎	▄▇▄▄▄▆▄█▄▁▁▁ beats 120/246	hard · avg 16%	▇ 0.4%	blunt
Decision Index 0.4.1: HLE (accuracy) submitted	14.2% ✎	▁▁▂▃▆█▇▅▅▃▁▁ beats 179/246	very hard · avg 13%	▇ 0.0%	blunt

vision▁·▃▂▃▃▇▄▄█▄▃ #44 of 79 · 1.9% of its score · 3 tests

Beats Surogate Rune 26B-A4B v3, Clef-omni, Sieve-Vision-4B here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
JevalKit: VisualDecisionBench images	81.8% ✎	▄▄···█▄·█▄█▄ beats 5/10	easy · avg 78%	▇ 1.0%	blunt
Benchmark Heaven imagejevbench v0.3.0: intelligence (points)	16.3059 points ✎	▃·▃▃··▅█▅▅·▃ beats 1/12	avg 32.235 points	▇ 0.6%	blunt
Benchmark Heaven imagejevbench v0.3.0: calibration (points)	79.2565 points ✎	▃···▃·▅█▅▃▃▅ beats 10/12	avg 68.584 points	▇ 0.3%	blunt

workflows▁▁▁▂▁▂▂▆█▄▁▁ #65 of 114 · 10% of its score · 5 tests

Beats Clef-omni, Clef-flash, Bespoke Nimble 9B v3 here, though they rank higher overall. Trails InternLM Intern-Decision Qwen3.5-4B, StartLux-Decision v1.1 0.8B, Kev 4B r10 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
JevalKit: Procedural Typed Decisions	65.1% ✎	█·███▄·▄▄▄·▄ beats 5/12	easy · avg 67%	▇▇ 4.4%	sharp · at its level
System One Bench: Typed decisions (English)	64.8% ✎	▁▁·▁▁▁▅▄▇██▁ beats 8/29	medium · avg 63%	▇ 2.2%	blunt · at its level
JevalKit: DecisionBench standard eval	72.9% ✎	▃▃·▃···▅█▅·█ beats 3/12	easy · avg 70%	▇ 1.9%	at its level
JevalKit: JevBench public 231	74.9% ✎	█▄▄▄▄▄██▄··▄ beats 2/12	easy · avg 82%	▇ 1.1%	typical
JevalKit: Typed Decisions held-out modal labels	65.4% ✎	▃▅··▃·▃▃·█▃█ beats 1/12	easy · avg 74%	▇ 0.1%	blunt · at its level

other tests5 not in a task · 8.7% of its score

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: private same skills (equated points)	40.6500 points ✎	▂▂▂▃▃▃▄▅▅▇█▃ beats 99/247	avg 41.206 points	▇ 3.4%	blunt
JevalKit: JevAlt-Bench EN/TR/DE	84.9% ✎	▅··▃▃▅▃█··▅▃ beats 2/12	very easy · avg 87%	▇ 2.5%	typical
Decision Index 0.4.1: private new domains (equated points)	16.8100 points ✎	▃█▃▃▂▆▃▃▄▄▄▄ beats 63/247	avg 38.861 points	▇ 1.9%	blunt
JevalKit: Decision Index public 0.3 (official weighted points)	48.1300 points ✎	▃▃·▃·▃▃▃▅▅·█ beats 5/12	avg 39.649 points	▇ 0.8%	blunt
JevalKit: Chinese decision validation v0.2	82.0% ✎	▃····█·█▃▃█▃ beats 0/12	very easy · avg 88%	▇ 0.1%	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/LiquidAI ↗

$ how the score was worked out · Decision Index 0.4.1 69%, System One Bench (Ali Alhashi… 18%, JevalKit (Valen Team) 12% and 1 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	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 69%	CC BY 4.0
System One Bench (Ali Alhashimi)	16	▇▇▇▇▇ 18%	CC-BY-4.0
JevalKit (Valen Team)	8	▇▇▇▇ 12%	Apache-2.0
Benchmark Heaven (own decision evaluations)	2	▇ 1%	CC BY 4.0
prior · pull toward the middle		▇ 0%
weighted average		= 50.9

not counted (2)

Decision Index 0.4.1	Decision Index 0.4.1: RouterBench (selected quality (quality objective)) [correct]	79.5%	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]	69.2%	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