Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.6 Terra | Kimi K3 |
|---|---|---|
| GPQA Diamond | 92.9 | 93.5 |
| Humanity's Last Exam | — | 56 |
coding
| Benchmark | GPT-5.6 Terra | Kimi K3 |
|---|---|---|
| SWE-Bench Pro | 63.4 | — |
multimodal
| Benchmark | GPT-5.6 Terra | Kimi K3 |
|---|---|---|
| MMMU-Pro | 80.7 | 81.6 |
| CharXiv Reasoning | — | 91.3 |
agent
| Benchmark | GPT-5.6 Terra | Kimi K3 |
|---|---|---|
| BrowseComp | 87.5 | 91.2 |
| Toolathlon | 53.1 | 73.2 |
| MCP Atlas | — | 84.2 |
| APEX Agents | — | 37.6 |
math
| Benchmark | GPT-5.6 Terra | Kimi K3 |
|---|---|---|
| FrontierMath | 84.9 | — |
general
| Benchmark | GPT-5.6 Terra | Kimi K3 |
|---|---|---|
| MRCR v2 | 89.6 | — |
Category Winners
coding
77.6
math
97.4
reasoning
M
Kimi K391.8
general
97.0
multimodal
M
Kimi K387.0
agent
M
Kimi K393.2