Atlas

Compare Models

Claude Sonnet 5.5 vs MiniMax M3.1 Flash

Claude Sonnet 5.5 leads by 6.8 capability points (≈95% interval +4.1 to +9.4): a clear lead over MiniMax M3.1 Flash · P(higher) ≈ 100%

Wins 24 of 26 shared benchmarks (2 lost)

Capability vs. relative cost
Claude Sonnet 5.5MiniMax 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.

Claude Sonnet 5.5vsMiniMax M3.1 Flashcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins24 – 2 · 26 shared

Comparing matching setups only (57 benchmarks excluded) · suites counted once
← Claude Sonnet 5.5MiniMax M3.1 Flash →

Claude Sonnet 5.5vsMiniMax M3.1 Flashwins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Neel Nanda and 16 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 Sonnet 5.5.

Toloka Arena (v3)Agents64.022.0−42.0
ALE-BenchCode24481495−33.8
Visa Agentic Commerce BenchmarkAgents64.433.3−31.1
NoCoT-Bench NCKI (kspine v3)General QA10765.9−28.5
FrontierML ($200 wallet)Agents50.423.2−27.2
GBENCH One-shot Coding GScoreGames67.441.1−26.2
o11y-bench 1.1 Pass^3Agents89.164.1−25.0
BuseyBench SVGMultimodal6.54.2−23.1
DAXBench v2Code86.263.8−22.3
Vibe Code Bench v1.1Code92.473.9−18.4
TaxCalcBench TY25Finance & Accounting18.00.0−18.0
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.036.1−17.9
APEX-Agents 1.1Professional Work83.768.1−15.6
NoCoT-Bench NCRI 15.2General QA95.168.4−14.7
APEX-1 (Mercor, mean score)Professional Work53.965.0+11.1
Humanity's Last Exam DiamondGeneral QA37.528.1−9.4
LiveBench 2.0Agents82.573.1−9.4
Design Arena Overall Code CategoriesCode13491297−8.6
Excel Modeling Benchmark v1.1Finance & Accounting75.967.8−8.1
APEX-Accounting 1.1Finance & Accounting64.560.0−4.5
GDPxProfessional Work8.24.0−4.2
HLAB-Vals-1.1Legal10.47.5−2.9
SAGE v1.1Math88.085.3−2.8
Agents' Last Exam v1.1 ScoreAgents60.457.9−2.5
Access to Justice BenchmarkLegal74.973.3−1.6
SWE-Bench Pro V2 — FullCode95.896.6+0.8

Claude Sonnet 5.5vsMiniMax M3.1 Flashnormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Neel Nanda and 16 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…