Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude Sonnet 4 | Gemini 2.5 Pro Preview 06-05 |
|---|---|---|
| GPQA Diamond | 75.4 | 86.4 |
| Humanity's Last Exam | — | 21.6 |
math
| Benchmark | Claude Sonnet 4 | Gemini 2.5 Pro Preview 06-05 |
|---|---|---|
| AIME 2025 | 70.5 | 88 |
coding
| Benchmark | Claude Sonnet 4 | Gemini 2.5 Pro Preview 06-05 |
|---|---|---|
| SWE-bench Verified | 72.7 | 67.2 |
| TerminalBench | 35.5 | — |
multimodal
| Benchmark | Claude Sonnet 4 | Gemini 2.5 Pro Preview 06-05 |
|---|---|---|
| MMMU | 74.4 | 82 |
agent
| Benchmark | Claude Sonnet 4 | Gemini 2.5 Pro Preview 06-05 |
|---|---|---|
| TAU-Bench Retail | 80.5 | — |
Category Winners
coding
59.7
math
66.8
reasoning
63.7
general
63.2
multimodal
83.8
agent
83.4