Claude Haiku 5.5 vs MiniMax M3.1 Flash
Claude Haiku 5.5 leads by 1.5 capability points (≈95% interval −0.9 to +3.8): statistically tied · P(higher) ≈ 89%
Wins 24 of 33 shared benchmarks (9 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.
Claude Haiku 5.5vsMiniMax M3.1 Flashcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev
Benchmark wins24 – 9 · 33 shared
Claude Haiku 5.5vsMiniMax M3.1 Flashwins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis 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 Claude Haiku 5.5.
| o11y-bench 1.1 Pass^3 | Agents | 85.9 | 64.1 | −21.9 |
| DAXBench v2 | Code | 81.9 | 63.8 | −18.1 |
| ALE-Bench | Code | 1996 | 1495 | −17.7 |
| Vibe Code Bench v1.1 | Code | 90.4 | 73.9 | −16.5 |
| GBENCH One-shot Coding GScore | Games | 55.6 | 41.1 | −14.5 |
| NoCoT-Bench NCKI (kspine v3) | General QA | 85.9 | 65.9 | −13.9 |
| FrontierML ($200 wallet) | Agents | 34.9 | 23.2 | −11.7 |
| Terminal-Bench 4.0 | Code | 36.4 | 25.3 | −11.1 |
| Legal Research Bench | Legal | 43.8 | 34.1 | −9.6 |
| SWE-Bench Pro V2 — Full | Code | 87.2 | 96.6 | +9.3 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Legal | 4.9 | 11.9 | +7.0 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Agents | 43.0 | 36.1 | −6.9 |
| GDPval-AA v2.1 | Professional Work | 1620 | 1465 | −6.8 |
| Finance Agent v2.1 | Finance & Accounting | 65.4 | 58.8 | −6.5 |
| AA-Briefcase v1.1 Elo | Agents | 1578 | 1474 | −5.7 |
| Vals Index v2.2 | Indexes | 58.7 | 53.8 | −4.9 |
| SAGE v1.1 | Math | 89.8 | 85.3 | −4.6 |
| BuseyBench SVG | Multimodal | 4.6 | 4.2 | −3.4 |
| HLAB-Vals-1.1 | Legal | 4.2 | 7.5 | +3.3 |
| Vals Legal Index | Indexes | 70.3 | 67.1 | −3.2 |
| Design Arena Overall Code Categories | Code | 1280 | 1297 | +2.8 |
| LiveBench 2.0 | Agents | 75.6 | 73.1 | −2.5 |
| Excel Modeling Benchmark v1.1 | Finance & Accounting | 65.5 | 67.8 | +2.4 |
| APEX-Accounting 1.1 | Finance & Accounting | 62.2 | 60.0 | −2.2 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Code | 50.2 | 48.2 | −2.0 |
| TaxCalcBench TY25 | Finance & Accounting | 2.0 | 0.0 | −2.0 |
| Vals Finance Index | Indexes | 57.4 | 55.9 | −1.5 |
| Toloka Arena (v3) | Agents | 21.0 | 22.0 | +1.0 |
| GDPx | Professional Work | 4.9 | 4.0 | −0.9 |
| Tax Agent Bench | Finance & Accounting | 62.2 | 62.7 | +0.5 |
| APEX-1 (Mercor, mean score) | Professional Work | 64.7 | 65.0 | +0.3 |
| APEX-Agents 1.1 | Professional Work | 68.3 | 68.1 | −0.2 |
| NoCoT-Bench NCRI 15.2 | General QA | 68.3 | 68.4 | 0.0 |
Claude Haiku 5.5vsMiniMax M3.1 Flashnormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis and 13 othersVillage LLM Atlasatlas.village.dev