Atlas

Compare Models

Claude Opus 5 vs Qwen3.8 Max (0902)

Claude Opus 5 leads by 3.8 capability points (≈95% interval +1.3 to +6.2): a clear lead over Qwen3.8 Max (0902) · P(higher) ≈ 100%

Wins 32 of 36 shared benchmarks (4 lost)

Capability vs. relative cost
Claude Opus 5Qwen3.8 Max (0902)Closed weightsOpen weights

Each line runs through one model’s priced reasoning settings, cheapest to most expensive; grey points are other models’ priced settings in the same range. Results whose reasoning effort the source didn’t report are left out wherever a model has named effort levels. Capability and relative cost are the published estimates (relative cost is cost per task as a multiple of GPT-5 (high)); hover a setting, or use the arrow keys, to set it beside the other model’s same setting.

Claude Opus 5vsQwen3.8 Max (0902)capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins32 – 4 · 36 shared

Comparing matching setups only (62 benchmarks excluded) · suites counted once
← Claude Opus 5Qwen3.8 Max (0902) →

Claude Opus 5vsQwen3.8 Max (0902)wins by categoryData Nov 15, 2026Data: Artificial Analysis, AIMING Lab (UNC-Chapel Hill), BuseyBench and 8 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Qwen3.8 Max (0902) minus Claude Opus 5.

Medical Long Context Reasoning (MLCR-AA)Health & Medicine59.420.0−39.4
FrontierSWE V2Code52.017.8−34.2
Terminal-Bench 4.0Code53.927.0−27.0
Humanity's Last Exam DiamondGeneral QA44.824.0−20.8
CritPt v2 (agentic, avg@5)Science71.451.4−20.0
Terminal-Bench-Science 0.1Science28.611.9−16.7
AA-Omniscience IndexIndexes37.112.0−12.5
CritPt (avg@5)Science29.117.7−11.4
Humanity's Last Exam (Text-Only)General QA54.443.1−11.3
Artificial Analysis Intelligence Index v5.0.1Indexes44.233.7−10.5
Terminal-Bench 4.0 (Artificial Analysis)Code49.038.9−10.1
AA-AnalystAgent (pass^5)Professional Work53.845.0−8.7
GDPval-AA v2Professional Work18611689−8.5
FrontierCode 2.0 (Artificial Analysis)Code31.223.4−7.7
RSI-Exam 0.1 (Full 88)Agents0.50.4−7.2
Epoch Capabilities Index (ECI)Indexes164156−7.0
AA-Briefcase EloAgents17201622−5.7
Artificial Analysis Intelligence Index v4.3.2Indexes50.845.4−5.4
Artificial Analysis Intelligence Index v4.3Indexes50.745.4−5.3
BuseyBench SVGMultimodal6.35.8−5.1
SciCode 1.0.1 (Artificial Analysis)Code56.452.1−4.3
Artificial Analysis Progress Index v1Indexes334324−3.5
DAXBench v2Code86.283.0−3.2
𝜏³-BankingAgents44.747.8+3.1
SPU (Science Prediction Universe)Science0.20.1−2.9
GDPval-AA v2.1Professional Work17291674−2.4
AA-LCR v1.1Search82.080.3−1.7
MMMU ProMultimodal84.082.8−1.2
GDP.pdf (Artificial Analysis, All-pass)Professional Work21.622.8+1.2
GPQA Diamond (Artificial Analysis avg@5)General QA93.792.8−0.9
Co-Design BenchCode66.165.3−0.8
Finance Agent v2Finance & Accounting58.657.9−0.7
AutomationBench-AA 1.0.6Agents56.656.2−0.4
Terminal-Bench 2.1Code89.188.8−0.3
Harvey LAB-AA v1.0 (criterion pass rate)Legal93.593.6+0.2
EnterpriseOps-Gym-AAAgents47.547.6+0.1

Claude Opus 5vsQwen3.8 Max (0902)normalised score marginsData Nov 15, 2026Data: Artificial Analysis, AIMING Lab (UNC-Chapel Hill), BuseyBench and 8 othersVillage LLM Atlasatlas.village.dev

Win counts use each benchmark’s declared score direction. Represented suite slices and the derived Capability score are excluded from headline totals. Missing or incompatible setups are excluded, never counted as losses.
Loading Atlas data…