Atlas

Compare Models

DeepSeek V4.1 Flash vs MiniMax M3.1 Flash

MiniMax M3.1 Flash leads by 0.5 capability points (≈95% interval −1.8 to +2.7): statistically tied · P(higher) ≈ 65%

DeepSeek V4.1 Flash wins 17 of 28 shared benchmarks (11 lost)

Capability vs. relative cost
DeepSeek V4.1 FlashMiniMax M3.1 FlashClosed 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 FlashvsMiniMax M3.1 Flashcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins17 – 11 · 28 shared

Comparing matching setups only (45 benchmarks excluded) · suites counted once
← DeepSeek V4.1 FlashMiniMax M3.1 Flash →

DeepSeek V4.1 FlashvsMiniMax M3.1 Flashwins by categoryData Nov 15, 2026Data: Vals AI, Mercor, DeepSeek and 13 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is MiniMax M3.1 Flash minus DeepSeek V4.1 Flash.

RSI Arena Stage 1 human ratingAgents8871002+55.3
APEX-Agents 1.1Professional Work50.568.1+17.6
FoodTruck BenchGames5017963464+17.1
Hebbia Financial Services BenchmarkFinance & Accounting63.147.0−16.1
ALE-BenchCode18291495−11.8
DAXBench v2Code75.563.8−11.7
Vibe Code Bench v1.1Code84.773.9−10.8
APEX-Accounting 1.1Finance & Accounting50.160.0+9.9
Humanity's Last Exam (Text-Only)General QA39.148.0+8.9
Legal Research BenchLegal41.334.1−7.2
GBENCH One-shot Coding GScoreGames48.041.1−6.8
Finance Agent v2.1Finance & Accounting65.458.8−6.6
Terminal-Bench 4.0Code19.725.3+5.6
LiveBench 2.0Agents78.073.1−5.0
NL2Repo - release-reported repository generation scoreCode65.461.2−4.2
TaxCalcBench TY25Finance & Accounting4.00.0−4.0
VoxelBench Text RatingGames18611810−3.6
Design Arena Overall Code CategoriesCode13161297−3.2
τ³-Banking (Mercor run, bm25_grep retrieval)Agents37.836.1−1.7
BabyVision with toolsMultimodal89.691.0+1.4
BuseyBench SVGMultimodal4.14.2+1.3
Vals Finance IndexIndexes57.155.9−1.2
APEX-1 (Mercor, mean score)Professional Work65.565.0−0.5
DeepSWE v1.1Code74.273.7−0.5
Access to Justice BenchmarkLegal72.973.3+0.4
Tax Agent BenchFinance & Accounting62.562.7+0.3
SAGE v1.1Math85.485.3−0.1
Excel Modeling Benchmark v1.1Finance & Accounting67.867.8+0.1

DeepSeek V4.1 FlashvsMiniMax M3.1 Flashnormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, DeepSeek and 13 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…