whatmodel.sh / decision / InternLM Intern-Decision Qwen3.5-2B

InternLM Intern-Decision Qwen3.5-2B

+ contribute

InternLM Intern-Decision Qwen3.5-2B is an open-weights decision model from Shanghai AI Laboratory (InternLM). Its published size is 2.27B parameters. Licence: apache-2.0.

Catalogue identity notes

Exact Decision Index tested engine intern-decision-2b; reported model InternLM Intern-Decision Qwen3.5-2B. Type: full fine-tune; technique: autoregressive. Implementation: https://github.com/internlm/Intern-Decision. 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.

36.7±0.9#278 of 355 decision modelsfrom 64 benchmark results across 4 sources

$ how it compares

The decision models ranked either side of InternLM Intern-Decision Qwen3.5-2B, out of 357 with a score, then more from Shanghai AI Laboratory (InternLM).

#	model	score	where	released
278	mojev (DecisionBench 0c8695b6) MoLeMo-Lab	37±2	request access
279	Crivo-400M tardellirs	37±1	model files
280	InternLM Intern-Decision Qwen3.5-2B Shanghai AI Laboratory (InternLM)	37±1	model files
281	Sieve 2B Sthānika AI	37±1	model files	2026-10-13
282	smalljev-semantic-v9 (DecisionBench 8e2bac7c) isHeSatoshi	36±2	model files
	more from Shanghai AI Laboratory (InternLM)
222	InternLM Intern-Decision Qwen3.5-4B Shanghai AI Laboratory (InternLM)	48±1	model files
310	InternLM Intern-Decision Qwen3.5-0.8B Shanghai AI Laboratory (InternLM)	30±1	model files

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

$ what it's good at

tools & automation▁▃▁▁▃▂▃▄▆██▆ #215 of 250 · 7.2% of its score · 6 tests

Beats Cam1 here, though they rank higher overall. Trails Prumo (rule-based engine), JevK5 Lite, Crivo-32M v2 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: BFCL (case exact accuracy) submitted	71.0% ✎	▁▂▁▁▃▂▂▄▅▆█▁ beats 34/246	very easy · avg 92%	▇ 3.7%	sharp · noisy
Decision Index 0.4.1: API-Bank (accuracy) submitted	28.5% ✎	▁▁▁▁▁▁▃▂▃▂█▂ beats 22/236	easy · avg 70%	▇ 1.8%	sharp · at its level · noisy
Decision Index 0.4.1: When2Call (accuracy) submitted	49.1% ✎	▁▁▁▂▃▅▅▆█▇█▃ beats 40/246	easy · avg 72%	▇ 0.8%	at its level · noisy
Decision Index 0.4.1: ToolRet (nDCG@10) submitted	49.0% ✎	▁▁▁▁▁·▁▃▁▄█▁ beats 30/246	medium · avg 60%	▇ 0.6%	blunt · at its level
Decision Index 0.4.1: Home appliance simulator (case exact accuracy) submitted	1.1% ✎	█▁▁▃▄▄▆▄▁▂▁▃ beats 51/243	medium · avg 37%	▇ 0.3%	sharp · noisy
Decision Index 0.4.1: RouterBench (selected quality (quality objective)) submitted	79.3% ✎	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▁▁▁·▁▁▁▃▃▃█▅ #215 of 258 · 3.8% of its score · 7 tests

Beats openvons Qwen3-4B (decision readout), Intelif Qwen3-4B, mini-jev Qwen3-4B (decision readout) here, though they rank higher overall. Trails Bosun v3.1 0.6B LoRA, Crivo-32M v2, Mola-68M here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: Amazon ESCI (macro-F1) submitted	34.6% ✎	▁▁▁▁▁▁▂▃▃▃█▅ beats 44/246	medium · avg 47%	▇ 1.2%	at its level
Decision Index 0.4.1: BRIGHT (nDCG@10) submitted	21.4% ✎	▁▁▁▁▁▃▁▂▂▄▇█ beats 22/246	medium · avg 38%	▇ 1.2%	blunt · at its level
Decision Index 0.4.1: CLINC150+OOS (macro-F1) submitted	4.3% ✎	▁▁▃▁▁▁▁▂▃▆█▄ beats 6/238	easy · avg 83%	▇ 0.4%	sharp · at its level · noisy
Decision Index 0.4.1: HoVer (accuracy) submitted	56.8% ✎	▆▆▇▅▄▄▃▇▄█▆▄ beats 32/246	easy · avg 74%	▇ 0.4%	blunt · at its level
Decision Index 0.4.1: PhishNChips phishing decisions (accuracy) submitted	70.0% ✎	█▅▅▄▃▂▁▁▁▁·▁ beats 129/245	easy · avg 70%	▇ 0.4%	at its level · noisy
Decision Index 0.4.1: BANKING77 (macro-F1) submitted	12.9% ✎	▁·▁▁▁▃▁▁▂█▇█ beats 6/238	easy · avg 79%	▇ 0.3%	at its level · noisy
Decision Index 0.4.1: SGD/SGD-X (macro-F1) submitted	59.3% ✎	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▂▂▂▃▂▅▆▃█▆▄▃ #211 of 256 · 19% of its score · 9 tests

Beats Valen-2B, Saina Helm 2 Omni (0.8B), Gevva E2B here, though they rank higher overall. Trails DecisionTune 1.0, vera-spark, Bekko System One v0 400M 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	60.4% ✎	▂▄▂▃▄▇▃▇█▅▃▁ beats 45/246	very easy · avg 85%	▇▇▇ 7.2%	sharp · at its level
Decision Index 0.4.1: NLI4CT (macro-F1) submitted	66.3% ✎	▁▁▁▂▂▃▃▄▇▇█▄ beats 63/246	easy · avg 74%	▇▇ 4.1%	at its level
Decision Index 0.4.1: ANLI (macro-F1) submitted	46.9% ✎	▁▂▃▅▅█▂····▁ beats 52/246	medium · avg 60%	▇ 3.3%	at its level
Decision Index 0.4.1: iSarcasmEval (Sarcasm F1 · track A, English) submitted	28.7% ✎	▁·▁·▁▁▂▄▅█▇▄ beats 55/246	medium · avg 44%	▇ 1.1%	at its level
Decision Index 0.4.1: ContractNLI (macro-F1) submitted	44.8% ✎	▁▁▁▁▁▁▃▂▂▇▇█ beats 29/246	easy · avg 71%	▇ 0.8%	at its level · noisy
Decision Index 0.4.1: FinEntity (macro-F1) submitted	63.2% ✎	▁···▁▁▁▃▃▆█▃ beats 28/246	easy · avg 84%	▇ 0.7%	at its level · noisy
Decision Index 0.4.1: VAST (macro-F1) submitted	39.3% ✎	▁▁▁▂▅▄▄▂▄▆█▄ beats 27/246	medium · avg 62%	▇ 0.6%	at its level
Decision Index 0.4.1: RAGTruth (F1 on hallucinated class) submitted	29.7% ✎	▁▁▁▁▁▁▁▃▁▃▆█ beats 30/246	easy · avg 66%	▇ 0.5%	at its level · noisy

1 more test

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: ACOS (per-review F1) submitted	4.0% ✎	▁··▂▃▃▅▇█▇▃▃ beats 19/246	hard · avg 22%	▇ 0.4%	noisy

arts & human taste▁▇▄▃▅▅█▅▄▂▁▁ #198 of 247 · 5.8% of its score · 6 tests

Beats vLLM-SR d3-nano (2B), Valen-4B, Saracura PT-BR 4B here, though they rank higher overall. Trails Qwen-2.5-1B-RLCD (decision readout), Bekko System One v0 68M, Crivo-17M 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	40.3% ✎	▁▁▂▃▃▃▇▅█▅▃▁ beats 36/246	medium · avg 61%	▇ 2.5%	at its level
Decision Index 0.4.1: BPoMP (accuracy) submitted	58.9% ✎	▁█▄▄▅█▆▅▄▁▁▁ beats 49/246	easy · avg 83%	▇ 1.3%	sharp · at its level · noisy
Decision Index 0.4.1: Humicroedit (accuracy) submitted	57.8% ✎	▁·▁▃▂▃▅█▄▂▁▁ beats 75/246	medium · avg 60%	▇ 1.1%	blunt · at its level
Decision Index 0.4.1: cfcolor (accuracy) submitted	55.8% ✎	▁········▁▇█ beats 57/246	medium · avg 60%	▇ 0.6%	blunt · at its level
Decision Index 0.4.1: Habermas Machine (accuracy) submitted	45.7% ✎	▁▂▄█▇▅▁▁▄▄▁▁ beats 143/246	medium · avg 47%	▇ 0.2%	blunt · at its level
Decision Index 0.4.1: POP909-CL (accuracy) submitted	0.9% ✎	▄▅▅▃▆▅▃▅▂▅▃█ beats 14/237	hard · avg 21%	▇ 0.2%	sharp · too hard for it · noisy

knowledge & reasoning▁▂▂▂▂▃▃▅▅▆█▂ #198 of 263 · 37% of its score · 14 tests

Beats OneJev1.6-0.8B, GLiClass v4 edge (0.4B), Crivo-1B here, though they rank higher overall. Trails laya (DecisionBench 1c5edc17), DM-JEPA 1.1, Sieve 2B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: ARC-Challenge (accuracy) submitted	78.2% ✎	▂▂▂▂▂▃▁▃▄▆█▆ beats 67/246	very easy · avg 90%	▇▇▇ 7.4%	sharp · at its level
Decision Index 0.4.1: ARC-Easy (accuracy) submitted	90.6% ✎	▁▁▂▁▂▂▃▂▃▆█▅ beats 68/246	very easy · avg 96%	▇▇▇ 6.5%	sharp · noisy
Decision Index 0.4.1: MMLU (accuracy) submitted	57.3% ✎	▁▄▄▃▃▂▅▇▅█▇▁ beats 70/246	easy · avg 70%	▇▇ 5.4%	sharp · at its level
Decision Index 0.4.1: MMLU-Pro (accuracy) submitted	33.0% ✎	▄▄▃▃▃▄▅▇██▅▁ beats 69/246	medium · avg 46%	▇▇ 4.7%	sharp · at its level
Decision Index 0.4.1: BBH (accuracy) submitted	47.3% ✎	▄▆▄▄█▇▇▄▃▁▁▁ beats 58/246	easy · avg 66%	▇▇ 4.4%	at its level
Decision Index 0.4.1: WinoGrande (accuracy) submitted	57.5% ✎	█▅█▅▄█▆▄▃▃▁▁ beats 39/246	easy · avg 78%	▇ 2.7%	at its level
Decision Index 0.4.1: GPQA Diamond (accuracy) submitted	34.2% ✎	▂▆▅▅▆█▆▃▁··▁ beats 82/246	medium · avg 39%	▇ 1.5%	blunt · at its level
Decision Index 0.4.1: GSM8K (accuracy (4- and 10-choice)) submitted	22.0% ✎	▁··▄█▆▆▆▅▅▃▁ beats 47/246	medium · avg 45%	▇ 1.5%	sharp · at its level

6 more tests

test	result	vs the others in its run	difficulty	counts	why
Decision Index 0.4.1: MuSR (accuracy) submitted	54.4% ✎	▁▁▂█▅▄▁▁▁▁▁▁ beats 58/246	medium · avg 60%	▇ 0.7%	blunt · at its level
Decision Index 0.4.1: CLadder (accuracy) submitted	53.8% ✎	▁███▆▅▃▃▁▁▁▁ beats 31/246	easy · avg 71%	▇ 0.7%	at its level
Decision Index 0.4.1: CRUXEval (accuracy) submitted	33.9% ✎	██▄▃▃▄▁▁···▁ beats 21/246	medium · avg 58%	▇ 0.6%	at its level · noisy
Decision Index 0.4.1: SATA-Bench (case exact accuracy) submitted	11.2% ✎	▁▁▁▁▁▁▃▂▅█▂▁ beats 48/246	hard · avg 22%	▇ 0.5%	typical
Decision Index 0.4.1: ChessBench (accuracy) submitted	10.6% ✎	▄▇▄▄▄▆▄█▄▁▁▁ beats 68/246	hard · avg 16%	▇ 0.3%	blunt
Decision Index 0.4.1: HLE (accuracy) submitted	12.8% ✎	▁▁▂▃▆█▇▅▅▃▁▁ beats 142/246	very hard · avg 13%	▇ 0.0%	blunt

vision▁·▃▂▃▃▇▄▄█▄▃ #68 of 79 · 7.9% of its score · 13 tests

Beats StartLux-Decision v1.1 0.8B, mmastrac diffgemma (decision readout), Winnow-E4B here, though they rank higher overall. Trails LiquidAI d1-omni-600M here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
JevalKit: VisualDecisionBench images	76.5% ✎	▄▄···█▄·█▄█▄ beats 2/10	easy · avg 78%	▇ 1.1%	blunt · at its level
Decision Index Vision 0.4 · Mind2Web (chance-corrected points) submitted	49.9800 points ✎	▁·▁▁▃▁▄▄▃▅█▂ beats 5/61	avg 63.423 points	▇ 0.9%	blunt
Decision Index Vision 0.4 · MMMU-Pro vision (chance-corrected points) submitted	15.8700 points ✎	▁▁▅▁▄▁▄█▄▃▇▇ beats 11/61	avg 22.797 points	▇ 0.8%	blunt
Decision Index Vision 0.4 · Winoground (chance-corrected points) submitted	63.2500 points ✎	▁·▁▁▁▁·▃▁▃▅█ beats 11/61	avg 66.987 points	▇ 0.7%	blunt
Decision Index Vision 0.4: private same skills (points)	42.7100 points ✎	▁▁▃▃▃▂▄▆▃▃▅█ beats 13/61	avg 48.067 points	▇ 0.7%	blunt
Decision Index Vision 0.4 · Moderation (Hateful Memes) (chance-corrected points) submitted	13.9300 points ✎	▆▃▃▆▄▄▇██▁▆▆ beats 6/61	avg 28.443 points	▇ 0.6%	blunt
Decision Index Vision 0.4 · RealWorldQA (chance-corrected points) submitted	50.3400 points ✎	▁·▂▂▁▃▃▅▅▃▂█ beats 12/61	avg 53.651 points	▇ 0.6%	blunt
Decision Index Vision 0.4 · CharXiv (chance-corrected points) submitted	58.1500 points ✎	▂▁▁▁▁▃▁▅▅▆██ beats 13/61	avg 62.137 points	▇ 0.6%	blunt

5 more tests

test	result	vs the others in its run	difficulty	counts	why
Decision Index Vision 0.4 · InfographicVQA (chance-corrected points) submitted	78.6900 points ✎	▁·▁▂▁▃▃▁▂▂▅█ beats 11/61	avg 83.859 points	▇ 0.4%	blunt
Decision Index Vision 0.4 · R-Bench-M (chance-corrected points) submitted	10.4800 points ✎	▂▄▇▆█▄▄▇▂··▁ beats 4/61	avg 19.676 points	▇ 0.4%	blunt
Decision Index Vision 0.4 · BLINK (chance-corrected points) submitted	42.9700 points ✎	▁···▁▂▃▂█▄▂▃ beats 13/61	avg 44.994 points	▇ 0.4%	blunt
Decision Index Vision 0.4 · CV-Bench (chance-corrected points) submitted	65.0300 points ✎	▁▁▁▁▂▃▁▁▄█▃▃ beats 12/61	avg 67.662 points	▇ 0.3%	blunt
Decision Index Vision 0.4 · KIE (CORD+FUNSD) (chance-corrected points) submitted	84.4400 points ✎	▁··▁▁▁▁▃▂▄█▁ beats 9/61	avg 85.184 points	▇ 0.2%	blunt

workflows▁▁▁▂▁▂▃▆█▄▁▁ #79 of 114 · 11% of its score · 6 tests

Beats vLLM-SR d3-lite (0.8B), Qwen3-Reranker-4B (DecisionBench 22e68366), Decision-1.0-Eos-0.8B (DecisionBench 3c2d6326) here, though they rank higher overall.

test	result	vs the others in its run	difficulty	counts	why
JevalKit: Procedural Typed Decisions	47.5% ✎	█·███▄·▄▄▄·▄ beats 1/12	easy · avg 67%	▇▇ 4.4%	sharp · at its level
Privacy policy · diagnostic43	55.8% ✎	█········▄·▄ beats 2/3	easy · avg 70%	▇ 2.6%	sharp · at its level
JevalKit: DecisionBench standard eval	55.2% ✎	▃▃·▃···▅█▅·█ beats 1/12	easy · avg 70%	▇ 1.9%	at its level
JevalKit: JevBench public 231	78.4% ✎	█▄▄▄▄▄██▄··▄ beats 3/12	easy · avg 82%	▇ 1.2%	at its level
Privacy policy · challenge56	100.0% ✎	▄·······▄··█ beats 2/3	very easy · avg 97%	▇ 0.6%	sharp · noisy
JevalKit: Typed Decisions held-out modal labels	79.3% ✎	▃▅··▃·▃▃·█▃█ beats 9/12	easy · avg 74%	▇ 0.1%	blunt · at its level

other tests5 not in a task · 8.5% of its score

test	result	vs the others in its run	difficulty	counts	why
JevalKit: JevAlt-Bench EN/TR/DE	78.6% ✎	▅··▃▃▅▃█··▅▃ beats 1/12	very easy · avg 87%	▇ 3.1%	typical
Decision Index 0.4.1: private same skills (equated points)	22.5000 points ✎	▂▂▂▃▃▃▄▅▅▇█▃ beats 43/247	avg 41.206 points	▇ 3.0%	blunt
Decision Index 0.4.1: private new domains (equated points)	18.5800 points ✎	▃█▃▃▂▆▃▃▄▄▄▄ beats 71/247	avg 38.861 points	▇ 1.6%	blunt
JevalKit: Decision Index public 0.3 (official weighted points)	15.0700 points ✎	▃▃·▃·▃▃▃▅▅·█ beats 1/12	avg 39.649 points	▇ 0.7%	blunt
JevalKit: Chinese decision validation v0.2	87.6% ✎	▃····█·█▃▃█▃ beats 3/12	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/internlm ↗

$ how the score was worked out · Decision Index 0.4.1 77%, JevalKit (Valen Team) 13%, Decision Index Vision 0.4 7% 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	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 77%	CC BY 4.0
JevalKit (Valen Team)	8	▇▇▇▇ 13%	Apache-2.0
Decision Index Vision 0.4 · submitted; independent execution unverified	12	▇▇ 7%	CC BY 4.0
FloorIsAwake · Privacy Decision Benchmark	2	▇ 3%	MIT
prior · pull toward the middle		▇ 0%
weighted average		= 36.7

not counted (2)

Decision Index 0.4.1	Decision Index 0.4.1: RouterBench (selected quality (quality objective)) [correct]	79.3%	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]	59.3%	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