d1-ultra

+ contribute

d1-ultra is a closed-weights decision model from Liquid AI, released 15 October 2026. Licence: proprietary.

Catalogue identity notes

System One Bench identifies the tested endpoint as liquid/d1-ultra. Hosted service snapshot, not an immutable weight revision. Separate from chat or open-weight models with similar names. API access verified 2026-10-31; benchmark dates are retained per result.

Model information sources

Release date ↗

69.3±1.6#21 of 355 decision modelsfrom 25 benchmark results across 4 sources

$ how it compares

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

#	model	score	where	released
19	MISA AI 2.0 Decision Plus · oak 1.1 (31B) MISA JSC	70±2	model files
20	vera-large scar-ai	70±2	model files
21	d1-ultra Liquid AI	69±2	api price unknown	2026-10-15
22	KnowLine 26B-A4B Gen2.5 PelaAI	69±2	model files	2026-10-22
23	OneJev1.6-27B OmniJev	69±2	model files	2026-10-17
	more from Liquid AI
32	d1.1 (hosted) Liquid AI	68±2	api price unknown
61	d1 (hosted) Liquid AI	65±1	api price unknown
198	LiquidAI d1-3B Liquid AI	51±1	model files

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

$ what it's good at

retrieval & classification▁▁▁·▁▁▁▁▃▃█▅ #46 of 258 · 13% of its score · 6 tests

Beats H2O Lightning 31B v1.1, AutoTrust-JEV-27B v1.1, MISA AI 2.0 Decision Plus · oak 1.1 (31B) here, though they rank higher overall. Trails StartLux-Decision v1.1 0.8B, Standard One 8B SH (v4), StartLux-Decision v1.1 2B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
System One Bench: MIRACL pairwise (Arabic)	91.0% ✎	▂···▁··▃▄▂█▇ beats 25/29	easy · avg 84%	▇▇ 4.4%	typical
System One Bench: MASSIVE intent (Arabic)	87.0% ✎	▄·▁·▁▃▃▃█▄▇▇ beats 19/29	easy · avg 76%	▇ 3.7%	typical
System One Bench: MIRACL rerank (Arabic)	89.0% ✎	▁▂·▁▁·▂·▄▄█▇ beats 23/29	easy · avg 79%	▇ 2.5%	typical
System One Bench: MASSIVE scenario (Arabic)	76.0% ✎	▃▁·▄▃▂█▃▁▁·▃ beats 24/29	medium · avg 60%	▇ 1.8%	typical
System One Bench: AG News (English)	91.0% ✎	▁···▅▁▅▂▆█▃▅ beats 24/29	easy · avg 88%	▇ 0.5%	blunt
System One Bench: Banking77 (English)	83.5% ✎	█▅▆█▅▃▃▃···▂ beats 6/26	very easy · avg 87%	▇ 0.0%	blunt · noisy

language understanding▂▂▂▃▂▅▆▃█▆▄▃ #21 of 256 · 12% of its score · 9 tests

Beats KnowLine 31B Gen1, Ainvest-Decision-27B v1.0, WeMM-JEV-27B here, though they rank higher overall. Trails Clef-flash, Clef-omni, Clef here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
System One Bench: BoolQ (English)	91.5% ✎	▁▁▁▃·▁▃▃▇▇█▅ beats 23/29	easy · avg 86%	▇ 3.1%	typical
System One Bench: OSACT4 offensive (Arabic)	94.5% ✎	▁······▁▃█▇▄ beats 25/29	easy · avg 84%	▇ 2.5%	sharp
System One Bench: SST-5 (English)	61.5% ✎	▃·▁▁▃·▇▄▄▄█▅ beats 26/29	medium · avg 51%	▇ 2.1%	blunt · at its level
System One Bench: SST-2 (English)	95.5% ✎	▁·▃▁▂▂▃▅█·▂▁ beats 20/29	easy · avg 92%	▇ 1.8%	typical
System One Bench: XNLI Arabic (Arabic)	76.0% ✎	▄▁▃▁▄▇█▅▃▁·▃ beats 18/29	medium · avg 68%	▇ 1.0%	at its level
System One Bench: XNLI French (French)	83.3% ✎	▂▅·▁▁▅▅█▃▁·▃ beats 22/29	medium · avg 76%	▇ 0.8%	blunt
System One Bench: Tweet sentiment (Arabic)	76.0% ✎	▁······▁▁▄█▄ beats 26/29	medium · avg 62%	▇ 0.8%	blunt · at its level
System One Bench: XNLI English (English)	92.7% ✎	▃·▁·▃▇█▄█▃▃▁ beats 25/29	easy · avg 85%	▇ 0.3%	blunt · noisy

1 more test

test	result	vs the others in its run	difficulty	counts	why
System One Bench: DAIR Emotion (English)	63.0% ✎	██▃·····▁··▁ beats 24/29	medium · avg 59%	▇ 0.0%	blunt · at its level

knowledge & reasoning▁▂▂▂▂▃▃▅▅▆█▃ #8 of 263 · 33% of its score · 3 tests

Beats Fastino GLiDE 1.2, Ainvest-Decision-27B v1.0, StartLux-Decision v1.1 27B here, though they rank higher overall. Trails Jev 1.13 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Etsabary · 1342 text decisions	80.8% ✎	▁··▂·▃▅▁▃▂▃█ beats 17/24	medium · avg 47%	▇▇▇▇▇▇▇▇ 19%	sharp
CaseHOLD · missing authority	73.6% ✎	▄···▄······█ beats 3/3	medium · avg 40%	▇▇▇ 7.6%	sharp · at its level
CONFLICTS · source conflict types	67.0% ✎	▃▃··▅▃···▃▃█ beats 9/9	medium · avg 45%	▇▇ 5.9%	at its level

vision▁·▃▂▃▃▇▄▄█▄▃ #1 of 79 · 37% of its score · 6 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.

test	result	vs the others in its run	difficulty	counts	why
Etsabary · 200 image decisions	74.5% ✎	▂▂··▂▃▂·█▃▂▂ beats 12/14	medium · avg 45%	▇▇▇▇▇▇ 14%	sharp
System One Bench: Arabic sentiment screenshot	62.5% ✎	▄··▄··▂···▆█ beats 8/11	very hard · avg 37%	▇▇ 5.4%	sharp · at its level · noisy
System One Bench: Image object count	91.5% ✎	▃····▃█▃▅·█▃ beats 7/11	easy · avg 71%	▇▇ 4.8%	sharp · maxed out
System One Bench: Visual yes/no questions	89.5% ✎	▂······▃█▃·▃ beats 11/11	medium · avg 80%	▇▇ 4.4%	typical
System One Bench: SST-2 screenshot	97.0% ✎	█···███·█·▄▄ beats 11/11	easy · avg 87%	▇▇ 4.4%	sharp
System One Bench: Oxford pets classification	87.5% ✎	███·█▄▄·▄··▄ beats 7/11	easy · avg 77%	▇ 3.6%	typical

workflows▁▁▁▂▁▂▂▆█▄▁▁ #17 of 114 * · 4.4% 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), Nace.AI Drex v2.0 (35B-A3B) here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
System One Bench: Typed decisions (English)	75.0% ✎	▁▁·▁▁▁▅▄▇██▁ beats 26/29	medium · avg 63%	▇▇ 4.4%	blunt

$ where to run it

Through an API

api · OpenRouter · price unknown · openrouter.ai ↗

Model repository and documentation: openrouter.ai ↗

$ how the score was worked out · System One Bench (Ali Alhashi… 52%, Etsabary · decision benchmarks 33%, Commit or Abstain? System One… 8% 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
System One Bench (Ali Alhashimi)	21	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 52%	CC-BY-4.0
Etsabary · decision benchmarks	2	▇▇▇▇▇▇▇▇▇▇ 33%	MIT
Commit or Abstain? System One Decision Models, LLMs, and Missing Legal Authority	1	▇▇ 8%	CC-BY-4.0
Do We Need an LLM to Know Why Sources Conflict? Decision Models for Knowledge-Conflict Typing in RAG	1	▇▇ 6%	CC-BY-4.0
prior · pull toward the middle		▇ 1%
weighted average		= 69.3

$ 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