Qwen3.5-Omni-Plus

Qwen3.5-Omni-Plus ranks first on 6 of the 52 benchmarks with five or more models scored.

Qwen Released March 30th, 2026 Family Qwen3.5-Omni The lab's model card Compare with others

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Qwen Mar 2026

Qwen3.5-Omni-Plus

Closed weights Flagship

142.9 #155 of 404

Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.

MMMU-Pro 73.9% #45 /90
IFBench 52.6% #63 /80
AA-LCR (Artificial Analysis Long Context Reasoning) 57% #27 /38

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

Benchmark	Score	Rank	Source
MMLU-Pro	85.9%	#20 of 167	Lab model card
MMLU-Redux	94.2%	#6 of 45	Lab model card
C-Eval	92%	#7 of 46	Lab model card
IFEval	89.7%	#34 of 135	Lab model card
SuperGPQA	66.4%	#10 of 35	Lab model card
LongBench v2	59.6%	#16 of 35	Lab model card
SimpleVQA	65.3%	#15 of 30	Lab model card
AA-LCR (Artificial Analysis Long Context Reasoning)	57%	#27 of 38	Lab model card
IFBench	52.6%	#63 of 80	Lab model card

Math and science

Benchmark	Score	Rank	Source
GPQA Diamond	83.9%	#89 of 302	Lab model card
HMMT Nov 2025	84.4%	#21 of 28	Lab model card
IMO-AnswerBench	65.5%	#36 of 37	Lab model card

Coding

Benchmark	Score		Rank	Source
LiveCodeBench v6	65.6%		#67 of 90	Lab model card

Agents and tools

Benchmark	Score	Rank	Source
tau2-bench	81%	#31 of 79	Lab model card
OmniGAIA	57.2%	#2 of 5	Official leaderboard
BFCL V4	63.3%	#17 of 30	Lab model card

Professional work

Benchmark	Score	Rank	Source
SLAKE	84.7%	#2 of 10	Lab model card
PMC-VQA	62.7%	#4 of 10	Lab model card
MedXpertQA-MM	54.7%	#11 of 21	Lab model card

Vision, audio and robotics

Benchmark	Score	Rank	Source
RefCOCO / RefCOCO+ / RefCOCOg	95%	#1 of 29	Lab model card
MVBench	79%	#1 of 23	Lab model card
EmbSpatialBench	85.4%	#1 of 16	Lab model card
MMAU	82.2%	#1 of 15	Lab model card
CC-OCR	83.4%	#1 of 14	Lab model card
VoiceBench	93.1%	#1 of 12	Lab model card
MMBench-EN v1.1	92.8%	#3 of 23	Lab model card
Daily-Omni	84.7%	#2 of 14	Official leaderboard
OCRBench	91.3%	#5 of 34	Lab model card
MathVista	86.1%	#12 of 70	Lab model card
MLVU	86.8%	#4 of 22	Lab model card
MMMU	80.1%	#22 of 103	Lab model card
WorldSense (omnimodal)	62.8%	#3 of 12	Official leaderboard
MMAR	80%	#1 of 4	Lab model card
ZeroBench	34.4%	#9 of 31	Lab model card
DynaMath	83.8%	#7 of 23	Lab model card
MMStar	79.4%	#11 of 36	Lab model card
ODinW-13	49.5%	#4 of 13	Lab model card
RealWorldQA	84.1%	#12 of 38	Lab model card
AI2D	91.2%	#17 of 52	Lab model card
MMSU	82.8%	#2 of 6	Lab model card
Omni-Cloze	64.8%	#1 of 3	Lab model card
WildSpeech-Bench	75.4	#1 of 3	Lab model card
VideoMME v2 (with subtitles)	27.4%	#11 of 31	Official leaderboard
CountBench	95.1%	#12 of 28	Lab model card
VoxMem	33%	#4 of 9	Official leaderboard
Video-MME	83.7%	#23 of 51	Lab model card
MMLongBench-Doc	57.5%	#9 of 19	Lab model card
MMMU-Pro	73.9%	#45 of 90	Lab model card
MuChoMusic	72.4%	#2 of 4	Lab model card
QIVD	68.5%	#2 of 4	Lab model card
SpeechRole	123.5	#1 of 2	Lab model card
MATH-Vision (MathVision)	73%	#29 of 51	Lab model card
τ-Voice	53.7%	#4 of 7	Official leaderboard
LVBench	71.2%	#19 of 31	Lab model card
ERQA	54.8%	#24 of 36	Lab model card
AV-SpeakerBench	71.3%	#2 of 3	Lab model card
AVUT	85%	#2 of 3	Lab model card
CharXiv	72.5%	#51 of 75	Lab model card
MMVU	67.5%	#19 of 26	Lab model card
Omni-InteractScope Explicit Request	13.9%	#3 of 4	Official leaderboard
Omni-InteractScope Proactive	31.1%	#3 of 4	Official leaderboard
FLEURS	6.55%	#5 of 6	Lab model card
MME-VideoOCR	77%	#1 of 1	Lab model card
URO-Bench	84.1%	#1 of 1	Lab model card

Its lab's numbers put it at 142.9, number 155 of 404

The Self-Reported Index fits one score to the 43 capability results Qwen reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 141.3 to 144.4. How we compute it.

CountBench 95.1
RefCOCO / RefCOCO+ / RefCOCOg 95
MMLU-Redux 94.2
VoiceBench 93.1
MMBench-EN v1.1 92.8
C-Eval 92
OCRBench 91.3
AI2D 91.2
IFEval 89.7
MLVU 86.8
MathVista 86.1
MMLU-Pro 85.9
EmbSpatialBench 85.4
SLAKE 84.7
HMMT Nov 2025 84.4
RealWorldQA 84.1
GPQA Diamond 83.9
DynaMath 83.8
Video-MME 83.7
CC-OCR 83.4
MMAU 82.2
tau2-bench 81
MMMU 80.1
MMStar 79.4
MVBench 79
MMMU-Pro 73.9
MATH-Vision (MathVision) 73
CharXiv 72.5
LVBench 71.2
MMVU 67.5
SuperGPQA 66.4
LiveCodeBench v6 65.6
IMO-AnswerBench 65.5
SimpleVQA 65.3
BFCL V4 63.3
PMC-VQA 62.7
LongBench v2 59.6
MMLongBench-Doc 57.5
AA-LCR (Artificial Analysis Long Context Reasoning) 57
MedXpertQA-MM 54.7
IFBench 52.6
ODinW-13 49.5
ZeroBench 34.4

Its lab's numbers next to 4 leaderboards

Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.

Benchmark	Lab says	Board says	Gap
Daily-Omni	84.6% DailyOmni	84.7% Leaderboard, Omni-Modal Language Models (With Visual and Audio)	Same
Omni-Cloze	64.8% Omni-Cloze	64.8% Leaderboard	Same
OmniGAIA	57.2% OmniGAIA (judged by deepseek-v3.2-thinking)	57.2% Leaderboard	Same
WorldSense (omnimodal)	62.8% WorldSense	62.8% Leaderboard, modality A+V	Same

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

Qwen3.5-Omni: Scaling Up, Toward Native Omni-Modal AGI (blog) Mar 30, 2026

DailyOmni 84.6
WorldSense 62.8
AVUT 85
AV-SpeakerBench 71.3
Best of VideoMME (with audio) 83.7 / VideoMME(w/o sub.) 81.9 83.7
QualcommInteractive 68.5
Omni-Cloze 64.8
OmniGAIA (judged by deepseek-v3.2-thinking) 57.2
MMAU 82.2
MMAR 80
MMSU 82.8
RUL-MuchoMusic 72.4
VoiceBench 93.1
Mean of URO-Bench-Pro U 66.3 / R 86.3 / O 99.8 84.1
SpeechRole 123.5
WildSpeech-Bench 75.4
Fleurs(top60) ASR WER (S2TT xx to zh/en BLEU 32.8) 6.55
Mean of CV15 zh 3.46 / yue 1.95 / zh-tw 2.27 / en 4.83 3.13
Mean of Librispeech test-clean 1.11 / test-other 2.23 1.67
Seed-en custom voice stability (Seed-zh 1.07, Seed-hard 6.24) 1.35
MMMU 80.1
MMMU-Pro 73.9
MathVision 73
Mathvista (mini) 86.1
DynaMath 83.8
Best of ZEROBench 5 / ZEROBench_sub 34.4 34.4
RealWorldQA 84.1
MMStar 79.4
MMBench EN-DEV-v1.1 92.8
SimpleVQA 65.3
CharXiv (RQ) 72.5
CC-OCR 83.4
AI2D_TEST 91.2
MMLongBench-Doc 57.5
OCRBench 91.3
ERQA 54.8
CountBench 95.1
RefCOCO(avg) 95
ODInW13 49.5
EmbSpatialBench 85.4
MLVU(M-Avg) 86.8
MVBench 79
LVBench 71.2
MMVU 67.5
MME-VideoOCR 77
SLAKE 84.7
PMC-VQA 62.7
MedXpertQA-MM 54.7
MMLU-Pro 85.9
MMLU-Redux 94.2
SuperGPQA 66.4
C-Eval 92
IFEval 89.7
IFBench 52.6
AA-LCR 57
LongBench v2 59.6
GPQA 83.9
LiveCodeBench v6 65.6
HMMT Nov 25 84.4
IMOAnswerBench 65.5
BFCL-V4 63.3
TAU2Bench 81
Mean of Wenetspeech net 4.30 / meeting 5.84 5.07

Its system card

Every result in the longer documents the lab published with it. 2 of them aren't in the launch post.

Qwen3.5-Omni Technical Report Apr 17, 2026

Capabilities

MMLU-Pro 85.9
MMLU-Redux 94.2
SuperGPQA 66.4
C-Eval 92
IFEval 89.7
IFBench 52.6
AA-LCR (Artificial Analysis Long Context Reasoning) 57
LongBench v2 59.6
GPQA Diamond 83.9
LiveCodeBench v6 65.6
HMMT Nov 2025 84.4
IMO-AnswerBench 65.5
BFCL V4 63.3
tau2-bench 81
MMAU 82.2
MMAR 80
MMSU 82.8
MuChoMusic 72.4
VoiceBench 93.1
URO-Bench
SpeechRole 123.5
WildSpeech-Bench 75.4
FLEURS 6.55
Common Voice
LibriSpeech
WenetSpeech
MMMU 80.1
MMMU-Pro 73.9
MATH-Vision (MathVision) 73
MathVista 86.1
DynaMath 83.8
ZeroBench 5
RealWorldQA 84.1
MMStar 79.4
MMBench-EN v1.1 92.8
SimpleVQA 65.3
CharXiv 72.5
CC-OCR 83.4
AI2D 91.2
MMLongBench-Doc 57.5
OCRBench 91.3
ERQA 54.8
CountBench 95.1
RefCOCO / RefCOCO+ / RefCOCOg 95
ODinW-13 49.5
EmbSpatialBench 85.4
Video-MME 83.7
MLVU 86.8
MVBench 79
LVBench 71.2
MMVU 67.5
MME-VideoOCR 77
Daily-Omni 84.6
WorldSense (omnimodal) 62.8
AVUT 85
AV-SpeakerBench 71.3
QIVD 68.5
Omni-Cloze 64.8
OmniGAIA 57.2
Seed-TTS-eval
MiniMax TTS Multilingual Test Set
CV3-Eval

Health

SLAKE 84.7
PMC-VQA 62.7
MedXpertQA-MM 54.7