Atlas

Compare Models

MiniMax M3.1 Flash vs Muse Spark 1.3

Muse Spark 1.3 leads by 1.3 capability points (≈95% interval −1.0 to +3.7): statistically tied · P(higher) ≈ 87%

Wins 21 of 29 shared benchmarks (8 lost)

Capability vs. relative cost
MiniMax M3.1 FlashMuse Spark 1.3Closed 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.

MiniMax M3.1 FlashvsMuse Spark 1.3capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins8 – 21 · 29 shared

Comparing matching setups only (51 benchmarks excluded) · suites counted once
← MiniMax M3.1 FlashMuse Spark 1.3 →

MiniMax M3.1 FlashvsMuse Spark 1.3wins by categoryData Nov 15, 2026Data: Mercor, Vals AI, Scale AI and 15 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Spark 1.3 minus MiniMax M3.1 Flash.

NoCoT-Bench NCKI (kspine v3)General QA65.9113+32.4
Visa Agentic Commerce BenchmarkAgents33.364.4+31.1
RSI Arena Stage 1 human ratingAgents1002938−30.8
DAXBench v2Code63.885.1+21.3
Legal Research BenchLegal34.155.3+21.2
ALE-BenchCode14951981+17.2
Professional Reasoning Benchmark - LegalLegal45.161.6+16.5
NoCoT-Bench NCRI 15.2General QA68.498.3+16.4
τ³-Banking (Mercor run, bm25_grep retrieval)Agents36.152.2+16.1
BuseyBench SVGMultimodal4.25.7+14.9
Professional Reasoning Benchmark - FinanceFinance & Accounting45.659.5+13.9
AutomationBench (Public Set)Agents62.849.6−13.2
Vibe Code Bench v1.1Code73.985.9+11.9
GBENCH One-shot Coding GScoreGames41.152.7+11.6
Tax Agent BenchFinance & Accounting62.772.4+9.7
APEX-1 (Mercor, mean score)Professional Work65.073.0+8.0
FrontierML ($200 wallet)Agents23.216.0−7.3
APEX-Agents 1.1Professional Work68.161.3−6.8
TaxCalcBench TY25Finance & Accounting0.06.0+6.0
Agents' Last Exam (ALE)Agents27.032.2+5.3
APEX-Accounting 1.1Finance & Accounting60.064.9+4.9
ErdosBenchMath3.02.9−3.7
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal11.915.0+3.1
Humanity's Last Exam DiamondGeneral QA28.125.4−2.7
DeepSWE v1.1Code73.775.4+1.7
LiveBench 2.0Agents73.174.8+1.7
HiL-Bench (Human-in-Loop Benchmark)Agents65.766.9+1.2
Access to Justice BenchmarkLegal73.372.2−1.1
Terminal-Bench 4.0Code25.324.7−0.5

MiniMax M3.1 FlashvsMuse Spark 1.3normalised score marginsData Nov 15, 2026Data: Mercor, Vals AI, Scale AI and 15 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…