benchgap
Alibaba (Qwen) · model

Qwen3.8 Max benchmark scores

As of 2026-11-22, Qwen3.8 Max (Alibaba (Qwen)) has measured scores on 68 benchmarks and estimated scores on 120 more.

BenchmarkScoreSource
AA-GPQA Diamond91.0%estimated ± 2.7 pp, high confidence
AA-HLE42.4%estimated ± 3.9 pp, high confidence
AA-SciCode52.0%estimated ± 3.1 pp, high confidence
AIME2699.2%estimated ± 1.2 pp, low confidence
Artificial Analysis Intelligence Index34.2%estimated ± 5.6 pp, medium confidence
CritPt17.9%estimated ± 4.0 pp, high confidence
GDPval-AA52.5%estimated ± 5.6 pp, medium confidence
GPQA Diamond92.6%measured
HLE43.6%measured
IFBench82.8%measured
AA-LCR74.3%estimated ± 3.4 pp, medium confidence
LiveCodeBench v695.8%estimated ± 1.7 pp, medium confidence
AA-Omniscience Accuracy45.6%estimated ± 6.5 pp, medium confidence
SciCode52.5%estimated ± 2.9 pp, low confidence
Terminal-Bench 2.186.6%measured
BrowseComp82.9%estimated ± 1.9 pp, high confidence
HLE w/ tools56.2%measured
IFEval95.0%estimated ± 3.6 pp, high confidence
LongBench v266.3%measured
τ²-bench results98.0%estimated ± 3.1 pp, high confidence
AA Coding Index72.0%estimated ± 2.5 pp, high confidence
AA-MMMU-Pro81.4%estimated ± 1.5 pp, high confidence
APEX-Agents37.5%estimated ± 7.4 pp, low confidence
APEX-Agents-AA36.3%estimated ± 2.9 pp, medium confidence
DeepSearchQA91.3%estimated ± 7.9 pp, low confidence
SWE-bench Verified81.6%estimated ± 2.3 pp, high confidence
AutomationBench27.3%measured
CyberGym78.7%estimated ± 4.6 pp, medium confidence
JobBench53.4%measured
SkillsBench70.2%measured
SWE-bench Pro67.7%measured
WideResearch81.9%measured
AA Agentic Index48.5%estimated ± 3.3 pp, medium confidence
MMLU-Pro85.7%estimated ± 2.5 pp, low confidence
AA-IFBench80.4%estimated ± 7.4 pp, low confidence
FailureSensorIQ67.9%estimated ± 1.7 pp, high confidence
GPQA92.6%measured
MMLU95.2%estimated ± 1.2 pp, medium confidence
Vals GPQA Diamond93.7%measured
Vals LiveCodeBench87.9%measured
Vals MMLU-Pro88.6%measured
MMMLU90.7%estimated ± 1.1 pp, low confidence
MMMU88.2%estimated ± 1.7 pp, medium confidence
Chess Puzzles46.0%measured
AA MMLU-Pro88.4%estimated ± 1.1 pp, high confidence
Gert Labs55.7%measured
AA-AnalystAgent42.0%estimated ± 4.6 pp, high confidence
AA EnterpriseOps-Gym46.3%estimated ± 1.2 pp, medium confidence
AA Harvey LAB v1.089.7%estimated ± 1.3 pp, medium confidence
AA Tau3 Banking41.8%estimated ± 3.3 pp, medium confidence
AA Terminal-Bench 2.182.0%estimated ± 2.5 pp, high confidence
ApprenticeBench17.0%estimated ± 6.0 pp, low confidence
ARC-AGI-193.0%estimated ± 2.5 pp, high confidence
ARC-AGI-271.2%estimated ± 9.7 pp, medium confidence
BabaIsBench58.0%measured
CursorBench 3.159.6%estimated ± 5.6 pp, low confidence
CursorBench 3.264.2%estimated ± 3.3 pp, medium confidence
EBR-bench29.0%estimated ± 10.4 pp, low confidence
Terminal-Bench 3.010.9%estimated ± 6.9 pp, low confidence
FrontierCode 1.1 Main42.3%estimated ± 3.5 pp, low confidence
FrontierMath v2 (Tier 4)46.3%measured
FrontierMath v2 (Tiers 1-3)74.7%measured
FrontierSWE v215.8%measured
Furniture Assembly36.0%estimated ± 9.3 pp, medium confidence
MaxxBench64.8%estimated ± 5.7 pp, medium confidence
MazeBench0.0%measured
MirrorCode52.2%measured
Mystery Game Puzzles38.0%measured
NYT Connections88.3%measured
OSWorld-Verified86.1%measured
ParseBench-171.2%estimated ± 1.0 pp, low confidence
Terminal-Bench 4.123.5%estimated ± 10.9 pp, low confidence
Terminal-Bench Hard51.7%estimated ± 3.9 pp, high confidence
Vals SWE-bench85.6%measured
Terminal-Bench 2.1 (Vals AI)67.4%measured
VulcanBench v381.2%measured
CWE-Bench-AA59.7%estimated ± 4.0 pp, medium confidence
AA Cyber Index32.8%estimated ± 9.4 pp, low confidence
DeepsecBench-AA17.5%estimated ± 5.2 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)57.9%estimated ± 4.4 pp, low confidence
HealthBench Professional (raw, Artificial Analysis)64.9%estimated ± 2.6 pp, low confidence
AA ITBench46.1%estimated ± 4.6 pp, medium confidence
Bug Hunt Bench25.7%measured
CursorBench 4.036.7%estimated ± 2.6 pp, medium confidence
CWE-bench v157.4%estimated ± 5.1 pp, medium confidence
DeepSWE56.6%measured
DRACO86.9%estimated ± 0.7 pp, low confidence
FrontierCode 2.024.3%estimated ± 4.8 pp, medium confidence
FrontierPhysics11.8%estimated ± 4.7 pp, high confidence
HLE w/o tools43.6%measured
OSWorld 2.019.4%measured
OSWorld 2.169.8%estimated ± 3.6 pp, high confidence
OSWorld 2.1 Offline63.9%estimated ± 5.5 pp, low confidence
PostTrainBench v1.128.6%estimated ± 3.7 pp, high confidence
SpireClimb5.8%estimated ± 7.2 pp, medium confidence
StealthMark62.4%measured
Toolathlon-Verified72.5%measured
VulcanBench Frontier v477.1%estimated ± 5.0 pp, high confidence
AA AutomationBench41.9%estimated ± 7.9 pp, low confidence
GDP.pdf16.8%estimated ± 4.1 pp, medium confidence
AA Terminal-Bench 4.034.2%estimated ± 11.0 pp, low confidence
ARC-AGI-35.3%estimated ± 7.2 pp, medium confidence
cursorBench4138.4%estimated ± 4.3 pp, medium confidence
CWE-bench45.4%measured
MKA38.5%estimated ± 3.2 pp, high confidence
SkillsBench official leaderboard (BenchFlow)56.6%estimated ± 6.1 pp, medium confidence
BinaryAudit35.8%estimated ± 12.1 pp, low confidence
CompileBench100.0%estimated ± 10.0 pp, low confidence
OTelBench22.5%estimated ± 6.6 pp, medium confidence
FACTS Multimodal51.5%measured
Hermes Bench70.5%measured
Hermes Index36.2%measured
ParseBench-275.6%estimated ± 5.0 pp, medium confidence
CharXiv93.5%measured
CharXiv w/o tools88.4%measured
ExploitBench59.8%estimated ± 7.2 pp, medium confidence
ExploitGym11.9%estimated ± 6.3 pp, medium confidence
AI2D_TEST93.5%estimated ± 0.9 pp, low confidence
C-Eval93.2%estimated ± 1.2 pp, medium confidence
Claw-Eval72.8%estimated ± 4.5 pp, medium confidence
CountBench97.3%estimated ± 1.3 pp, low confidence
DynaMath88.5%estimated ± 0.8 pp, low confidence
ERQA77.8%measured
HMMT Feb 2026100.0%estimated ± 1.5 pp, low confidence
HMMT Nov 202595.6%estimated ± 1.5 pp, low confidence
MathVision95.2%measured
MCP Atlas85.5%estimated ± 6.6 pp, low confidence
MCP-Tasks84.2%estimated ± 6.0 pp, low confidence
MLVU (M-Avg)90.8%measured
MMAnswerBench92.9%estimated ± 2.4 pp, low confidence
MMLU-Redux95.1%estimated ± 1.7 pp, medium confidence
MMMU-Pro82.3%measured
NL2Repo55.9%measured
OmniDocBench 1.592.1%measured
PostTrainBench v131.3%estimated ± 3.6 pp, high confidence
RealWorldQA88.0%measured
ScreenSpot Pro84.5%measured
SimpleVQA75.0%measured
SuperGPQA83.2%estimated ± 7.2 pp, medium confidence
Terminal-Bench 2.071.3%estimated ± 5.8 pp, medium confidence
Toolathlon49.4%estimated ± 5.2 pp, low confidence
V*96.1%estimated ± 6.5 pp, low confidence
Video-MME (w/o subtitle)86.7%estimated ± 1.3 pp, medium confidence
Video-MME (with subtitle)90.4%measured
VideoMMMU88.7%measured
We-Math87.5%estimated ± 2.1 pp, low confidence
AA LiveCodeBench95.0%estimated ± 2.6 pp, medium confidence
HealthBench Hard33.4%estimated ± 5.1 pp, low confidence
LiveCodeBench Pro89.4%estimated ± 6.0 pp, low confidence
MedXpertQA (MM)80.4%measured
MedXpertQA (Text)62.4%estimated ± 4.5 pp, medium confidence
MMLU-Pro (Arcee)87.7%estimated ± 3.4 pp, low confidence
React Native Evals83.9%estimated ± 4.7 pp, low confidence
ResearchClawBench20.7%estimated ± 1.9 pp, medium confidence
PostTrainBench33.1%estimated ± 2.4 pp, medium confidence
DAYJOB13.6%estimated ± 3.2 pp, high confidence
DAYJOB Healthcare0.7%estimated ± 4.2 pp, high confidence
Finance Agent v256.4%estimated ± 2.4 pp, medium confidence
FrontierCode 1.1 Extended58.3%estimated ± 3.0 pp, medium confidence
HealthBench Professional52.7%estimated ± 4.2 pp, medium confidence
HLE-Verified54.3%estimated ± 4.6 pp, medium confidence
LABBench282.3%estimated ± 1.8 pp, low confidence
ScarfBench63.7%measured
BlenderWorld (text-only)73.2%estimated ± 2.8 pp, medium confidence
Cookie-Bench83.1%estimated ± 1.3 pp, medium confidence
SimpleQA38.6%estimated ± 13.4 pp, low confidence
Agents' Last Exam52.4%measured
BabyVision w/ Python91.3%measured
OpenHarmony Bench v1.060.8%measured
OpenHarmony Bench v1.267.8%measured
SEC-Bench Pro82.8%estimated ± 5.6 pp, medium confidence
ZeroBench w/ Python49.0%measured
BabyVision82.0%measured
MMVU82.4%measured
PerceptionBench63.5%measured
ZeroBench24.0%measured
MRCRv292.9%measured
LVBench81.8%measured
CC-OCR79.6%measured
RefCOCO (avg)86.7%estimated ± 1.8 pp, medium confidence
OCRBench V274.2%measured
SWE-Marathon46.5%estimated ± 2.1 pp, medium confidence
BioBasicsBench31.2%estimated ± 2.2 pp, medium confidence
LifeVal75.2%estimated ± 3.8 pp, medium confidence
MMMU-Pro w/ Python83.2%estimated ± 2.6 pp, high confidence
ExploitBench Internal Port1.0%estimated ± 1.4 pp, medium confidence
MathVision w/ Python97.7%measured
SWE-Atlas Codebase QnA53.7%estimated ± 4.8 pp, high confidence