Atlas

Compare Models

DeepSeek V4.1 Flash 1017 vs Qwen3.8 Max (0902)

DeepSeek V4.1 Flash 1017 leads by 0.6 capability points (≈95% interval −1.9 to +3.1): statistically tied · P(higher) ≈ 69%

Wins 24 of 44 shared benchmarks (20 lost)

Capability vs. relative cost
DeepSeek V4.1 Flash 1017Qwen3.8 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.

DeepSeek V4.1 Flash 1017vsQwen3.8 Max (0902)capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins24 – 20 · 44 shared

Comparing matching setups only (177 benchmarks excluded) · suites counted once
← DeepSeek V4.1 Flash 1017Qwen3.8 Max (0902) →

DeepSeek V4.1 Flash 1017vsQwen3.8 Max (0902)wins by categoryData Nov 15, 2026Data: Artificial Analysis, BuseyBench, DAXBench and 4 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 DeepSeek V4.1 Flash 1017.

AA-PersonalityBench Conviction indexChat & Writing94.335.3−59.6
AA-PersonalityBench Transparency indexChat & Writing86.744.8−42.3
AA-PersonalityBench Slop indexChat & Writing39.173.3+34.6
AA-PersonalityBench Informativeness indexChat & Writing71.439.1−32.7
AA-PersonalityBench Expressiveness indexChat & Writing52.423.8−28.8
AA-PersonalityBench Warmth indexChat & Writing27.748.6+21.2
AA-PersonalityBench Organization indexChat & Writing88.671.4−17.3
BuseyBench SVGMultimodal4.35.8+14.4
AutomationBench-AA 1.0.6Agents68.556.2−12.3
AA-PersonalityBench Imagination indexChat & Writing37.248.6+11.5
AA-Briefcase v1.1 EloAgents14161604+10.3
VIALS-AA (pass^3)Multimodal40.431.7−8.7
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine26.720.0−6.7
CritPt (avg@5)Science24.017.7−6.3
VIALS-AA (pass@3)Multimodal66.572.7+6.2
GDP.pdf (Artificial Analysis, All-pass)Professional Work17.222.8+5.6
DAXBench v2Code77.783.0+5.3
Terminal-Bench 4.0 (Artificial Analysis)Code33.838.9+5.1
FrontierCode 2.0 (Artificial Analysis)Code28.523.4−5.0
Artificial Analysis Intelligence Index v5.0Indexes28.533.6+5.0
Terminal-Bench 4.0Code31.027.0−4.0
Artificial Analysis Economics IndexIndexes38.542.5+4.0
SciCode 1.0.1 (Artificial Analysis)Code55.952.1−3.8
AA-Omniscience IndexIndexes4.712.0+3.6
Artificial Analysis Legal IndexIndexes37.741.0+3.3
Humanity's Last Exam DiamondGeneral QA20.824.0+3.2
MMMU ProMultimodal79.782.8+3.1
Artificial Analysis Finance & Accounting IndexIndexes42.245.0+2.9
Epoch Capabilities Index (ECI)Indexes159156−2.8
AA-LCR v1.1Search82.780.3−2.3
VIALS-AA (average of 3)Multimodal53.250.9−2.3
Visa Agentic Commerce BenchmarkAgents57.860.0+2.2
Artificial Analysis Strategy & Ops IndexIndexes49.247.3−2.0
Artificial Analysis Engineering IndexIndexes37.639.3+1.7
TCSExamMath46.044.8−1.2
MKA (Artificial Analysis)Professional Work38.537.4−1.1
Artificial Analysis Intelligence Index v5.0.1Indexes32.633.7+1.1
Terminal-Bench-Science 0.1Science12.911.9−1.0
CritPt v2 (agentic, avg@5)Science52.351.4−0.9
FrontierCode 1.1 Main (Artificial Analysis)Code43.743.2−0.5
Artificial Analysis Healthcare & Medical IndexIndexes41.340.8−0.5
Artificial Analysis Intelligence Index v4.3.2Indexes45.145.4+0.3
Humanity's Last Exam (Text-Only)General QA43.343.1−0.2
Artificial Analysis Progress Index v1Indexes3243240.0

DeepSeek V4.1 Flash 1017vsQwen3.8 Max (0902)normalised score marginsData Nov 15, 2026Data: Artificial Analysis, BuseyBench, DAXBench and 4 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…