Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT OSS 20B | MiniMax M1 80K |
|---|---|---|
| GPQA Diamond | 71.5 | 70 |
| Humanity's Last Exam | 10.9 | 8.4 |
agent
| Benchmark | GPT OSS 20B | MiniMax M1 80K |
|---|---|---|
| TAU-Bench Retail | 54.8 | 63.5 |
math
| Benchmark | GPT OSS 20B | MiniMax M1 80K |
|---|---|---|
| AIME 2025 | — | 76.9 |
coding
| Benchmark | GPT OSS 20B | MiniMax M1 80K |
|---|---|---|
| SWE-bench Verified | — | 56 |
general
| Benchmark | GPT OSS 20B | MiniMax M1 80K |
|---|---|---|
| SimpleQA | — | 18.5 |
Category Winners
coding
27.2
math
45.5
reasoning
34.4
general
22.4
agent
37.2