Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| GPQA Diamond | 79.9 | 89.5 |
| Humanity's Last Exam | 19.8 | 31.6 |
| ARC-AGI v2 | — | 40.1 |
math
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| AIME 2025 | 89.3 | — |
coding
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| SWE-bench Verified | 67.8 | 80.2 |
| TerminalBench | 37.7 | — |
| SWE-Bench Pro | — | 55.9 |
| SciCode | — | 48.7 |
general
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| SimpleQA | 97.1 | — |
agent
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| BrowseComp | 40.1 | 77.4 |
| Toolathlon | — | 54.4 |
| MCP Atlas | — | 79.6 |
multimodal
| Benchmark | DeepSeek-V3.2-Exp | Inkling-Small |
|---|---|---|
| MMMU-Pro | — | 74 |
| CharXiv Reasoning | — | 77.4 |
Category Winners
coding
T
Inkling-Small66.5
math
68.8
reasoning
T
Inkling-Small58.1
general
98.9
multimodal
T
Inkling-Small61.9
agent
T
Inkling-Small67.5