Atlas

Compare Models

GPT-6.1 Terra vs MiniMax M3.1 Flash

GPT-6.1 Terra leads by 5.4 capability points (≈95% interval +2.9 to +8.0): a clear lead over MiniMax M3.1 Flash · P(higher) ≈ 100%

Wins 24 of 32 shared benchmarks (8 lost)

Capability vs. relative cost
GPT-6.1 TerraMiniMax 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.

GPT-6.1 TerravsMiniMax M3.1 Flashcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins24 – 8 · 32 shared

Comparing matching setups only (40 benchmarks excluded) · suites counted once
← GPT-6.1 TerraMiniMax M3.1 Flash →

GPT-6.1 TerravsMiniMax M3.1 Flashwins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Neel Nanda and 14 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 GPT-6.1 Terra.

ALE-BenchCode28751495−48.9
GBENCH One-shot Coding GScoreGames77.641.1−36.4
NoCoT-Bench NCRI 15.2General QA13068.4−34.1
Toloka Arena (v3)Agents55.022.0−33.0
BuseyBench SVGMultimodal7.34.2−30.6
NoCoT-Bench NCKI (kspine v3)General QA11065.9−30.3
DAXBench v2Code93.663.8−29.8
Terminal-Bench 4.0Code49.525.3−24.2
TaxCalcBench TY25Finance & Accounting24.00.0−24.0
Code Migration v1.1 (Vals Index v2.2 subset)Code72.248.2−24.0
Visa Agentic Commerce BenchmarkAgents51.133.3−17.8
Agents' Last Exam v1.1 (ALE v1.1)Agents41.829.5−12.3
APEX-Agents 1.1Professional Work79.368.1−11.2
Vibe Code Bench v1.1Code84.073.9−10.1
Vals Legal IndexIndexes58.867.1+8.3
Legal Research BenchLegal26.034.1+8.2
Tax Agent BenchFinance & Accounting55.062.7+7.7
Humanity's Last Exam DiamondGeneral QA35.728.1−7.6
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal4.511.9+7.4
Vals Index v2.2Indexes60.553.8−6.8
LiveBench 2.0Agents79.773.1−6.6
HLAB-Vals-1.1Legal1.77.5+5.8
Excel Modeling Benchmark v1.1Finance & Accounting73.367.8−5.5
Agents' Last Exam Overall ScoreAgents60.055.0−5.0
GDPxProfessional Work8.54.0−4.5
APEX-Accounting 1.1Finance & Accounting63.760.0−3.7
Finance Agent v2.1Finance & Accounting56.058.8+2.9
Vals Finance IndexIndexes53.355.9+2.6
HiL-Bench (Human-in-Loop Benchmark)Agents68.065.7−2.3
τ³-Banking (Mercor run, bm25_grep retrieval)Agents37.136.1−1.0
APEX-1 (Mercor, mean score)Professional Work64.265.0+0.8
Design Arena Overall Code CategoriesCode13001297−0.5

GPT-6.1 TerravsMiniMax M3.1 Flashnormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Neel Nanda and 14 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…