whatmodel.sh / decision / Perplexity Decider v1.2 (27B)

Perplexity Decider v1.2 (27B)

+ contribute

Perplexity Decider v1.2 (27B) is an open-weights decision model from Perplexity. Its published size is 26.09B parameters. Licence: apache-2.0.

Catalogue identity notes

Exact Decision Index tested engine pplx-decider-v1.2; reported model Perplexity Decider v1.2 (27B). 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.

69.8±1.0#17 of 355 decision modelsfrom 79 benchmark results across 6 sources

$ how it compares

The decision models ranked either side of Perplexity Decider v1.2 (27B), out of 357 with a score, then more from Perplexity.

#	model	score	where	RAM	released
15	Bespoke Nimble v4 Max (27B) Bespoke Labs	70±1	model files		2026-10
16	Turion Jev 31B Turion.AI	70±1	model files
17	Perplexity Decider v1.2 (27B) Perplexity	70±1	model files
18	AutoTrust-JEV-27B v1.1 AutoTrust AI Lab	70±1	model files		2026-10-17
19	MISA AI 2.0 Decision Plus · oak 1.1 (31B) MISA JSC	70±2	model files
	more from Perplexity
10	Perplexity Decider v1.1 (27B) Perplexity	72±1	api price unknown		2026-10-05
90	Perplexity Decider Flash v1.2 (4B) Perplexity	62±1	local	~10GB

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

$ what it's good at

tools & automation▁▃▁▁▁▂▃▄▆██▆ #30 of 250 · 2.6% of its score · 6 tests

Beats Jeffan-31b-1710, Fastino GLiDE 1.2, Ainvest-Decision-27B v1.0 here, though they rank higher overall. Trails Solomon v1.1, Clef-flash, Clef-omni 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	87.4% ✎	▁▁▁▁▁▁▁▂▃▂█▂ beats 212/236	easy · avg 70%	▇ 0.8%	sharp · noisy
Decision Index 0.4.1: ToolRet (nDCG@10) submitted	67.5% ✎	▁▁▁▁▁·▁▁▁▄█▁ beats 191/246	medium · avg 60%	▇ 0.6%	blunt · at its level
Decision Index 0.4.1: When2Call (accuracy) submitted	90.3% ✎	▁▁▁▂▃▅▅▆█▇█▃ beats 221/246	easy · avg 72%	▇ 0.4%	noisy
Decision Index 0.4.1: Home appliance simulator (case exact accuracy) submitted	90.9% ✎	█▁▁▃▄▄▆▄▃▂▁▃ beats 207/243	medium · avg 37%	▇ 0.4%	sharp · noisy
Decision Index 0.4.1: BFCL (case exact accuracy) submitted	97.8% ✎	▁▂▁▁▁▂▂▄▅▆█▁ beats 217/246	very easy · avg 92%	▇ 0.4%	sharp · maxed out · noisy
Decision Index 0.4.1: RouterBench (selected quality (quality objective)) submitted	79.8% ✎	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▁▁▁·▁▁▁▁▃▃█▅ #9 of 258 · 10% of its score · 13 tests

Beats Fastino GLiDE 1.2, KnowLine 31B Gen1, Ainvest-Decision-27B v1.0 here, though they rank higher overall. Trails OneShot 27B, StartLux-Decision v1.1 9B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
System One Bench: MIRACL pairwise (Arabic)	91.5% ✎	▂···▁··▃▄▂█▇ beats 29/29	easy · avg 84%	▇ 1.9%	typical
Decision Index 0.4.1: BRIGHT (nDCG@10) submitted	50.1% ✎	▁▁▁▁▁▁▁▂▂▄▇█ beats 235/246	medium · avg 38%	▇ 1.7%	blunt · at its level
System One Bench: MASSIVE intent (Arabic)	91.0% ✎	▄·▁·▁▃▃▃█▄▇▇ beats 28/29	easy · avg 76%	▇ 1.6%	typical
Decision Index 0.4.1: Amazon ESCI (macro-F1) submitted	62.4% ✎	▁▁▁▁▁▁▂▂▃▃█▅ beats 236/246	medium · avg 47%	▇ 1.4%	at its level
System One Bench: MIRACL rerank (Arabic)	86.5% ✎	▁▂·▁▁·▂·▄▄█▇ beats 13/29	easy · avg 79%	▇ 1.1%	typical
System One Bench: MASSIVE scenario (Arabic)	79.0% ✎	▃▁·▄▃▂█▂▃▁·▃ beats 25/29	medium · avg 60%	▇ 0.8%	typical
Decision Index 0.4.1: HoVer (accuracy) submitted	86.3% ✎	▆▆▇▅▄▄▃▇▄█▆▄ beats 199/246	easy · avg 74%	▇ 0.3%	blunt
Decision Index 0.4.1: PhishNChips phishing decisions (accuracy) submitted	74.7% ✎	█▅▅▄▃▂▁▁▁▁·▁ beats 150/245	easy · avg 70%	▇ 0.3%	noisy

5 more tests

test	result	vs the others in its run	difficulty	counts	why
System One Bench: AG News (English)	91.5% ✎	▁···▅▁▅▂▆█▂▅ beats 27/29	easy · avg 88%	▇ 0.2%	blunt
Decision Index 0.4.1: BANKING77 (macro-F1) submitted	91.0% ✎	▁·▁▁▁▁▁▁▂█▇█ beats 199/238	easy · avg 79%	▇ 0.1%	noisy
Decision Index 0.4.1: CLINC150+OOS (macro-F1) submitted	94.4% ✎	▁▁▁▁▁▁▁▂▃▆█▄ beats 198/238	easy · avg 83%	▇ 0.1%	sharp · maxed out · noisy
System One Bench: Banking77 (English)	88.0% ✎	█▅▆█▅▃▃▃···▂ beats 14/26	very easy · avg 87%	▇ 0.0%	blunt · noisy
Decision Index 0.4.1: SGD/SGD-X (macro-F1) submitted	36.3% ✎	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▂▂▂▃▂▅▆▃█▆▄▃ #29 of 256 · 16% of its score · 18 tests

Beats KnowLine 31B Gen1, Ainvest-Decision-27B v1.0, H2O Lightning 31B v1.1 here, though they rank higher overall. Trails Clef-flash, Clef-omni, JEV-35B 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	75.5% ✎	▁▂▃▅▅█▂····▁ beats 198/246	medium · avg 60%	▇ 2.7%	typical
Decision Index 0.4.1: NLI4CT (macro-F1) submitted	85.7% ✎	▁▁▁▂▂▃▃▄▇▇█▄ beats 221/246	easy · avg 74%	▇ 2.5%	typical
Decision Index 0.4.1: iSarcasmEval (Sarcasm F1 · track A, English) submitted	74.7% ✎	▁·▁·▁▁▂▄▅█▇▄ beats 245/246	medium · avg 44%	▇ 1.4%	at its level
System One Bench: BoolQ (English)	91.5% ✎	▁▁▁▃·▁▃▃▇▇█▅ beats 25/29	easy · avg 86%	▇ 1.4%	typical
Decision Index 0.4.1: HellaSwag (accuracy) submitted	96.6% ✎	▂▄▂▂▄▇▃▇█▅▃▁ beats 216/246	very easy · avg 85%	▇ 1.2%	sharp · maxed out
System One Bench: OSACT4 offensive (Arabic)	95.0% ✎	▁······▁▃█▇▄ beats 27/29	easy · avg 84%	▇ 1.1%	sharp
Decision Index 0.4.1: ACOS (per-review F1) submitted	40.9% ✎	▁··▂▃▃▅▇█▇▃▃ beats 203/246	hard · avg 22%	▇ 1.0%	at its level · noisy
System One Bench: SST-5 (English)	60.0% ✎	▃·▁▁▃·▇▄▄▄█▅ beats 24/29	medium · avg 51%	▇ 0.9%	blunt · at its level

10 more tests

test	result	vs the others in its run	difficulty	counts	why
System One Bench: SST-2 (English)	95.5% ✎	▁·▃▁▂▂▃▅█·▂▁ beats 24/29	easy · avg 92%	▇ 0.8%	maxed out
Decision Index 0.4.1: VAST (macro-F1) submitted	80.2% ✎	▁▁▁▂▅▄▄▂▄▆█▄ beats 224/246	medium · avg 62%	▇ 0.5%	typical
Decision Index 0.4.1: ContractNLI (macro-F1) submitted	84.0% ✎	▁▁▁▁▁▁▁▂▂▇▇█ beats 206/246	easy · avg 71%	▇ 0.5%	noisy
System One Bench: XNLI Arabic (Arabic)	90.0% ✎	▄▁▃▁▄▇█▅▃▁·▃ beats 29/29	medium · avg 68%	▇ 0.4%	at its level
System One Bench: XNLI French (French)	90.0% ✎	▂▅·▁▁▅▅█▃▁·▃ beats 28/29	medium · avg 76%	▇ 0.4%	blunt
Decision Index 0.4.1: RAGTruth (F1 on hallucinated class) submitted	84.4% ✎	▁▁▁▁▁▁▁▁▁▃▆█ beats 228/246	easy · avg 66%	▇ 0.3%	noisy
System One Bench: Tweet sentiment (Arabic)	75.5% ✎	▁······▁▁▄█▄ beats 24/29	medium · avg 62%	▇ 0.3%	blunt · at its level
Decision Index 0.4.1: FinEntity (macro-F1) submitted	94.3% ✎	▁···▁▁▁▁▃▆█▃ beats 224/246	easy · avg 84%	▇ 0.2%	maxed out · noisy
System One Bench: XNLI English (English)	94.0% ✎	▃·▁·▃▇█▄█▃▃▁ beats 28/29	easy · avg 85%	▇ 0.1%	blunt · noisy
System One Bench: DAIR Emotion (English)	60.0% ✎	██▃·····▁··▁ beats 22/29	medium · avg 59%	▇ 0.0%	blunt · at its level

arts & human taste▁▇▄▃▅▅█▅▄▂▁▁ #54 of 247 · 5.1% of its score · 6 tests

Trails Kas 4B, baselabs joey-1 (2B), Clef-flash 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	79.0% ✎	▁▁▂▃▃▃▇▅█▅▃▁ beats 210/246	medium · avg 61%	▇ 2.0%	typical
Decision Index 0.4.1: Humicroedit (accuracy) submitted	65.8% ✎	▁·▁▃▂▂▅█▄▂▁▁ beats 214/246	medium · avg 60%	▇ 1.2%	blunt · at its level
Decision Index 0.4.1: cfcolor (accuracy) submitted	68.2% ✎	▁········▁▇█ beats 233/246	medium · avg 60%	▇ 0.7%	blunt · at its level
Decision Index 0.4.1: POP909-CL (accuracy) submitted	61.6% ✎	▄▅▅▃▆▅▃▅▂▅▃█ beats 178/237	hard · avg 21%	▇ 0.7%	sharp · at its level · noisy
Decision Index 0.4.1: BPoMP (accuracy) submitted	95.1% ✎	▁█▄▄▅█▆▅▄▁▁▁ beats 193/246	easy · avg 83%	▇ 0.4%	sharp · maxed out · noisy
Decision Index 0.4.1: Habermas Machine (accuracy) submitted	58.9% ✎	▁▂▄█▇▅▁▁▄▄▁▁ beats 194/246	medium · avg 47%	▇ 0.2%	blunt · at its level

knowledge & reasoning▁▂▂▂▂▃▃▅▅▆█▂ #22 of 263 · 35% of its score · 17 tests

Beats KnowLine 35B-A3B Gen1.5, Bespoke Nimble v4 Max (27B) here, though they rank higher overall. Trails d1 (hosted), Jev 1.13, Matilda Decision v1 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Etsabary · 1342 text decisions	83.5% ✎	▁··▂·▃▅▁▃▂▂█ beats 19/24	medium · avg 47%	▇▇▇ 8.2%	sharp
Decision Index 0.4.1: MMLU-Pro (accuracy) submitted	71.9% ✎	▄▄▃▃▂▄▅▇██▅▁ beats 219/246	medium · avg 46%	▇▇ 5.5%	sharp · at its level
CaseHOLD · missing authority	73.1% ✎	▄···▄······█ beats 2/3	medium · avg 40%	▇ 3.3%	sharp · at its level
Decision Index 0.4.1: BBH (accuracy) submitted	85.2% ✎	▄▆▄▄█▇▇▄▃▁▁▁ beats 215/246	easy · avg 66%	▇ 2.8%	typical
Decision Index 0.4.1: MMLU (accuracy) submitted	88.0% ✎	▁▄▄▃▃▂▅▇▅█▇▁ beats 221/246	easy · avg 70%	▇ 2.7%	sharp
CONFLICTS · source conflict types	66.7% ✎	▃▃··▅▃···▃▃█ beats 8/9	medium · avg 45%	▇ 2.6%	at its level
Decision Index 0.4.1: GPQA Diamond (accuracy) submitted	55.1% ✎	▂▆▅▅▆█▆▃▁··▁ beats 233/246	medium · avg 39%	▇ 2.2%	blunt · at its level
Decision Index 0.4.1: GSM8K (accuracy (4- and 10-choice)) submitted	83.7% ✎	▁··▄█▆▆▆▅▅▃▁ beats 230/246	medium · avg 45%	▇ 1.8%	sharp

9 more tests

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: WinoGrande (accuracy) submitted	94.0% ✎	█▅█▅▄█▆▄▃▃▁▁ beats 226/246	easy · avg 78%	▇ 1.3%	typical
Decision Index 0.4.1: SATA-Bench (case exact accuracy) submitted	23.2% ✎	▁▁▁▁▁▁▁▂▅█▂▁ beats 92/246	hard · avg 22%	▇ 1.2%	at its level
Decision Index 0.4.1: MuSR (accuracy) submitted	70.7% ✎	▁▁▂█▅▄▁▁▁▁▁▁ beats 222/246	medium · avg 60%	▇ 0.7%	blunt · at its level
Decision Index 0.4.1: ARC-Challenge (accuracy) submitted	97.1% ✎	▂▂▂▂▂▂▁▃▄▆█▆ beats 202/246	very easy · avg 90%	▇ 0.7%	sharp · maxed out
Decision Index 0.4.1: CRUXEval (accuracy) submitted	86.0% ✎	██▄▃▃▄▁▁···▁ beats 218/246	medium · avg 58%	▇ 0.6%	noisy
Decision Index 0.4.1: ChessBench (accuracy) submitted	24.6% ✎	▄▇▄▄▄▆▄█▄▁▁▁ beats 217/246	hard · avg 16%	▇ 0.6%	blunt
Decision Index 0.4.1: CLadder (accuracy) submitted	90.8% ✎	▁███▆▅▃▃▁▁▁▁ beats 222/246	easy · avg 71%	▇ 0.5%	typical
Decision Index 0.4.1: ARC-Easy (accuracy) submitted	99.4% ✎	▁▁▂▁▂▂▂▂▃▆█▅ beats 242/246	very easy · avg 96%	▇ 0.3%	sharp · maxed out · noisy
Decision Index 0.4.1: HLE (accuracy) submitted	16.0% ✎	▁▁▂▃▆█▇▅▅▃▁▁ beats 218/246	very hard · avg 13%	▇ 0.0%	blunt

vision▁·▃▂▃▃▇▄▄█▄▃ #4 of 79 · 24% of its score · 18 tests

Beats Fastino GLiDE 1.2, StartLux-Decision v1.1 27B, vLLM-SR d3-ultra (Decision 3.0, 27B) here, though they rank higher overall. Trails Jev 1.13, d1.1 (hosted), d1-ultra here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Etsabary · 200 image decisions	76.0% ✎	▂▂··▂▃▂·█▃▃▂ beats 13/14	medium · avg 45%	▇▇ 6.1%	sharp
System One Bench: Arabic sentiment screenshot	57.0% ✎	▄··▄··▂···▆█ beats 5/11	very hard · avg 37%	▇ 2.4%	sharp · at its level · noisy
System One Bench: Image object count	95.5% ✎	▃····▃█▃▅·█▃ beats 11/11	easy · avg 71%	▇ 2.0%	sharp · maxed out
System One Bench: Visual yes/no questions	86.5% ✎	▂······▃█▃·▃ beats 1/11	medium · avg 80%	▇ 1.9%	typical
System One Bench: SST-2 screenshot	94.5% ✎	█···███·█·▄▄ beats 6/11	easy · avg 87%	▇ 1.9%	sharp
System One Bench: Oxford pets classification	85.5% ✎	███·█▄▄·▄··▄ beats 5/11	easy · avg 77%	▇ 1.6%	typical
Decision Index Vision 0.4 · Mind2Web (chance-corrected points) submitted	93.2900 points ✎	▁·▁▁▃▁▄▄▃▅█▃ beats 60/61	avg 63.423 points	▇ 1.1%	blunt
Decision Index Vision 0.4 · MMMU-Pro vision (chance-corrected points) submitted	43.3800 points ✎	▁▁▅▁▄▁▄█▄▃▇▇ beats 48/61	avg 22.797 points	▇ 1.0%	blunt

10 more tests

test	result	vs the others in its run	difficulty	counts	why
Decision Index Vision 0.4 · Winoground (chance-corrected points) submitted	87.7500 points ✎	▁·▁▁▁▁·▂▁▃▅█ beats 61/61	avg 66.987 points	▇ 0.9%	blunt
Decision Index Vision 0.4: private same skills (points)	66.1800 points ✎	▁▁▃▃▁▂▄▆▃▃▅█ beats 50/61	avg 48.067 points	▇ 0.9%	blunt
Decision Index Vision 0.4 · Moderation (Hateful Memes) (chance-corrected points) submitted	55.3200 points ✎	▆▃▃▆▄▄▇██▁▆▆ beats 58/61	avg 28.443 points	▇ 0.8%	blunt
Decision Index Vision 0.4 · RealWorldQA (chance-corrected points) submitted	73.6400 points ✎	▁·▂▂▁▃▃▅▅▃▂█ beats 55/61	avg 53.651 points	▇ 0.7%	blunt
Decision Index Vision 0.4 · CharXiv (chance-corrected points) submitted	83.4600 points ✎	▂▁▁▁▁▃▁▅▅▆██ beats 60/61	avg 62.137 points	▇ 0.7%	blunt
Decision Index Vision 0.4 · InfographicVQA (chance-corrected points) submitted	98.3600 points ✎	▁·▁▂▁▁▃▁▂▂▅█ beats 58/61	avg 83.859 points	▇ 0.5%	blunt
Decision Index Vision 0.4 · R-Bench-M (chance-corrected points) submitted	34.4900 points ✎	▂▄▇▆█▄▄▇▂··▁ beats 48/61	avg 19.676 points	▇ 0.5%	blunt
Decision Index Vision 0.4 · BLINK (chance-corrected points) submitted	63.2700 points ✎	▁···▁▂▂▂█▄▂▃ beats 49/61	avg 44.994 points	▇ 0.5%	blunt
Decision Index Vision 0.4 · CV-Bench (chance-corrected points) submitted	77.5800 points ✎	▁▁▁▁▂▁▁▁▄█▃▃ beats 52/61	avg 67.662 points	▇ 0.4%	blunt
Decision Index Vision 0.4 · KIE (CORD+FUNSD) (chance-corrected points) submitted	92.4500 points ✎	▁··▁▁▁▁▁▂▄█▁ beats 45/61	avg 85.184 points	▇ 0.3%	blunt

workflows▁▁▁▂▁▂▂▆█▄▁▁ #14 of 114 * · 1.9% of its score · 1 test

Beats Nace.AI Drex Max v1 (27B) here, though they rank higher overall. Trails GPT-6 Luna Decisions, reflex-27b (DecisionBench 1d4bf0f2), Microsoft-Decision-1 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
System One Bench: Typed decisions (English)	75.1% ✎	▁▁·▁▁▁▅▄▇██▁ beats 27/29	medium · avg 63%	▇ 1.9%	blunt

other tests2 not in a task · 5.6% of its score

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: private same skills (equated points)	63.9900 points ✎	▂▂▂▃▃▃▄▅▅▇█▃ beats 235/247	avg 41.206 points	▇ 3.6%	blunt
Decision Index 0.4.1: private new domains (equated points)	73.9800 points ✎	▃█▃▃▂▆▃▃▄▄▄▄ beats 232/247	avg 38.861 points	▇ 2.0%	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/perplexity-ai ↗

$ how the score was worked out · Decision Index 0.4.1 48%, System One Bench (Ali Alhashi… 23%, Etsabary · decision benchmarks 14% 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	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 48%	CC BY 4.0
System One Bench (Ali Alhashimi)	21	▇▇▇▇▇▇▇ 23%	CC-BY-4.0
Etsabary · decision benchmarks	2	▇▇▇▇ 14%	MIT
Decision Index Vision 0.4 · submitted; independent execution unverified	12	▇▇▇ 8%	CC BY 4.0
Commit or Abstain? System One Decision Models, LLMs, and Missing Legal Authority	1	▇ 3%	CC-BY-4.0
Do We Need an LLM to Know Why Sources Conflict? Decision Models for Knowledge-Conflict Typing in RAG	1	▇ 3%	CC-BY-4.0
prior · pull toward the middle		▇ 0%
weighted average		= 69.8

not counted (2)

Decision Index 0.4.1	Decision Index 0.4.1: RouterBench (selected quality (quality objective)) [correct]	79.8%	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]	36.3%	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