Atlas

Compare Models

DeepSeek V4.1 Flash vs Qwen3.8-Max

Qwen3.8-Max leads by 0.3 capability points (≈95% interval −2.0 to +2.6): statistically tied · P(higher) ≈ 59%

Wins 18 of 32 shared benchmarks (13 lost, 1 tied)

Capability vs. relative cost
DeepSeek V4.1 FlashQwen3.8-MaxClosed 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 FlashvsQwen3.8-Maxcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins13 – 18 · 32 shared, 1 tied

Comparing matching setups only (257 benchmarks excluded) · suites counted once
← DeepSeek V4.1 FlashQwen3.8-Max →

DeepSeek V4.1 FlashvsQwen3.8-Maxwins by categoryData Nov 15, 2026Data: Artificial Analysis, Conceptual Reasoning Benchmark Authors, Alibaba and 10 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Qwen3.8-Max minus DeepSeek V4.1 Flash.

BabyVisionMultimodal43.382.0+38.7
AutomationBench-AA 1.0.6Agents68.949.2−19.7
DeepSWE v1.1Code74.256.6−17.6
GameBench 2Games68.451.4−17.0
YC-BenchGames16572362095838+15.3
BuseyBench SVGMultimodal4.15.4+13.3
NL2Repo - release-reported repository generation scoreCode65.455.9−9.5
Revelo Code IndexCode38.929.4−9.5
Terminal-Bench 2.1Code90.681.3−9.3
Terminal-Bench 4.0 (Artificial Analysis)Code26.818.7−8.1
GDP.pdf (Artificial Analysis, All-pass)Professional Work12.820.2+7.4
Humanity's Last ExamGeneral QA36.843.6+6.8
CritPt (avg@5)Science14.320.0+5.7
Agents' Last Exam - release-reported pass rate, subset unspecifiedAgents31.827.0−4.8
AA-Omniscience IndexIndexes-5.33.4+4.4
Nimble Web ContextSearch73.577.3+3.8
ACCoRDGeneral QA68.171.4+3.3
Medical Long Context Reasoning (MLCR-AA)Health & Medicine22.819.4−3.3
DTBenchGeneral QA89.992.0+2.1
AA-Briefcase EloAgents14241388−2.1
Dual Exam Agent IndexIndexes18.320.0+1.7
SciCode 1.0.1 (Artificial Analysis)Code51.953.2+1.4
Conceptual Reasoning IndexIndexes58.459.4+1.3
Epoch Capabilities Index (ECI)Indexes156157+0.8
Artificial Analysis Intelligence Index v4.3Indexes39.540.3+0.8
LMCAGeneral QA47.046.2−0.7
Artificial Analysis Intelligence Index v4.3.2Indexes39.540.2+0.7
GDPval-AA v2.1Professional Work16001611+0.5
Ramp Accounting BenchFinance & Accounting40.841.2+0.4
Artificial Analysis Progress Index v1Indexes320320−0.2
GDPval-AA v2Professional Work16321630−0.1
ZeroBench (Main questions, pass@5, with tools)Multimodal49.049.00.0

DeepSeek V4.1 FlashvsQwen3.8-Maxnormalised score marginsData Nov 15, 2026Data: Artificial Analysis, Conceptual Reasoning Benchmark Authors, Alibaba and 10 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…