Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT OSS 20B | MiniMax M1 40K |
|---|---|---|
| GPQA Diamond | 71.5 | 69.2 |
| Humanity's Last Exam | 10.9 | 7.2 |
agent
| Benchmark | GPT OSS 20B | MiniMax M1 40K |
|---|---|---|
| TAU-Bench Retail | 54.8 | 67.8 |
math
| Benchmark | GPT OSS 20B | MiniMax M1 40K |
|---|---|---|
| AIME 2025 | — | 74.6 |
coding
| Benchmark | GPT OSS 20B | MiniMax M1 40K |
|---|---|---|
| SWE-bench Verified | — | 55.6 |
general
| Benchmark | GPT OSS 20B | MiniMax M1 40K |
|---|---|---|
| SimpleQA | — | 17.9 |
Category Winners
coding
28.1
math
41.5
reasoning
36.5
general
21.7
agent
50.1