Grok 4.6 vs MiniMax M3.1 Flash
Grok 4.6 leads by 0.2 capability points (≈95% interval −2.2 to +2.5): statistically tied · P(higher) ≈ 55%
Wins 15 of 22 shared benchmarks (7 lost)
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.
Grok 4.6vsMiniMax M3.1 Flashcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev
Benchmark wins15 – 7 · 22 shared
Grok 4.6vsMiniMax M3.1 Flashwins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Neel Nanda 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 Grok 4.6.
| Toloka Arena (v3) | Agents | 56.0 | 22.0 | −34.0 |
| NoCoT-Bench NCKI (kspine v3) | General QA | 112 | 65.9 | −31.7 |
| DAXBench v2 | Code | 86.2 | 63.8 | −22.3 |
| GBENCH One-shot Coding GScore | Games | 55.4 | 41.1 | −14.2 |
| Legal Research Bench | Legal | 48.1 | 34.1 | −13.9 |
| ALE-Bench | Code | 1822 | 1495 | −11.6 |
| NoCoT-Bench NCRI 15.2 | General QA | 86.8 | 68.4 | −10.1 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Agents | 45.7 | 36.1 | −9.6 |
| APEX-Agents 1.1 | Professional Work | 77.3 | 68.1 | −9.2 |
| Tax Agent Bench | Finance & Accounting | 71.4 | 62.7 | −8.7 |
| Humanity's Last Exam Diamond | General QA | 19.6 | 28.1 | +8.5 |
| Terminal-Bench 4.0 | Code | 17.2 | 25.3 | +8.1 |
| FoodTruck Bench | Games | 59537 | 63464 | +5.0 |
| BuseyBench SVG | Multimodal | 4.6 | 4.2 | −3.9 |
| DeepSWE v1.1 | Code | 65.2 | 68.1 | +2.9 |
| APEX-1 (Mercor, mean score) | Professional Work | 67.7 | 65.0 | −2.7 |
| CyberGym | Cybersecurity | 79.7 | 82.2 | +2.5 |
| Vibe Code Bench v1.1 | Code | 76.2 | 73.9 | −2.3 |
| Access to Justice Benchmark | Legal | 71.3 | 73.3 | +2.0 |
| LiveBench 2.0 | Agents | 74.9 | 73.1 | −1.8 |
| Design Arena Overall Code Categories | Code | 1288 | 1297 | +1.5 |
| APEX-Accounting 1.1 | Finance & Accounting | 60.3 | 60.0 | −0.3 |
Grok 4.6vsMiniMax M3.1 Flashnormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Neel Nanda and 13 othersVillage LLM Atlasatlas.village.dev