Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-4.5 | GPT-4o mini |
|---|---|---|
| GPQA Diamond | 69.5 | 40.2 |
| BigBench-Hard | — | 78.5 |
| ARC-Challenge | — | 93.1 |
coding
| Benchmark | GPT-4.5 | GPT-4o mini |
|---|---|---|
| SWE-bench Verified | 38 | — |
| HumanEval | — | 87.2 |
multimodal
| Benchmark | GPT-4.5 | GPT-4o mini |
|---|---|---|
| MMMU | 75.2 | — |
| CharXiv Reasoning | 55.4 | — |
general
| Benchmark | GPT-4.5 | GPT-4o mini |
|---|---|---|
| SimpleQA | 62.5 | — |
| MMMLU | 85.1 | — |
| Chatbot Arena Elo | — | 1219 |
| MMLU-Pro | — | 63.2 |
| IFEval |
agent
| Benchmark | GPT-4.5 | GPT-4o mini |
|---|---|---|
| TAU-Bench Retail | 68.4 | — |
language
| Benchmark | GPT-4.5 | GPT-4o mini |
|---|---|---|
| HellaSwag | — | 91.2 |
| WinoGrande | — | 82.3 |
safety
| Benchmark | GPT-4.5 | GPT-4o mini |
|---|---|---|
| TruthfulQA | — | 65.8 |
Category Winners
coding
61.6
math
48.4
reasoning
52.8
general
70.3
language
35.0
multimodal
38.6
safety
62.9
agent
51.9