MiniMax-AI/MiniMax-M1 README — 'Performance of MiniMax-M1 on core benchmarks' (MiniMax-M1-80K column)
repo
Fields
id
sources/minimax-m1-github-readmeKind
repo
Accessed
2026-10-19
notes
Read https://github.com/MiniMax-AI/MiniMax-M1 for the benchmark figures on models/minimax-m1. As recorded by the reader: All values are the MiniMax-M1-80K column of the official README table (MiniMax-M1-40K alternatives: AIME 2024 83.3, AIME 2025 74.6, MATH-500 96.0, GPQA Diamond 69.2, MMLU-Pro 80.6, SWE-bench Verified 55.6). SWE-bench Verified uses the Agentless scaffold on n=486 verified tasks. README's LiveCodeBench row is labeled '(24/8~25/5)' (65.0), not v6, so it was excluded.
Notes
Read https://github.com/MiniMax-AI/MiniMax-M1 for the benchmark figures on models/minimax-m1. As recorded by the reader: All values are the MiniMax-M1-80K column of the official README table (MiniMax-M1-40K alternatives: AIME 2024 83.3, AIME 2025 74.6, MATH-500 96.0, GPQA Diamond 69.2, MMLU-Pro 80.6, SWE-bench Verified 55.6). SWE-bench Verified uses the Agentless scaffold on n=486 verified tasks. README's LiveCodeBench row is labeled '(24/8~25/5)' (65.0), not v6, so it was excluded.
Cross-references
This record references no other records.