Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-4 Turbo | MiniMax M2.5 |
|---|---|---|
| GPQA Diamond | 49.3 | — |
| BigBench-Hard | 85.1 | — |
| ARC-Challenge | 96.3 | — |
coding
| Benchmark | GPT-4 Turbo | MiniMax M2.5 |
|---|---|---|
| HumanEval | 87.6 | — |
| SWE-bench Verified | — | 80.2 |
| SWE-Bench Pro | — | 55.4 |
language
| Benchmark | GPT-4 Turbo | MiniMax M2.5 |
|---|---|---|
| HellaSwag | 95.3 | — |
| WinoGrande | 87.5 | — |
safety
| Benchmark | GPT-4 Turbo | MiniMax M2.5 |
|---|---|---|
| TruthfulQA | 70.1 | — |
general
| Benchmark | GPT-4 Turbo | MiniMax M2.5 |
|---|---|---|
| IFEval | 82.1 | — |
| MMLU-Pro | 64.1 | — |
| Chatbot Arena Elo | 1256 | — |
agent
| Benchmark | GPT-4 Turbo | MiniMax M2.5 |
|---|---|---|
| BrowseComp | — | 76.3 |
Category Winners
coding
63.4
math
60.3
reasoning
45.5
general
47.9
language
85.2
safety
84.8
agent
70.9