Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.1 High | Seed 2.1 Pro |
|---|---|---|
| GPQA Diamond | 88.1 | — |
| Humanity's Last Exam | — | 55.7 |
math
| Benchmark | GPT-5.1 High | Seed 2.1 Pro |
|---|---|---|
| AIME 2025 | 99.6 | — |
coding
| Benchmark | GPT-5.1 High | Seed 2.1 Pro |
|---|---|---|
| SWE-Bench Pro | — | 57.5 |
| SciCode | — | 59.8 |
multimodal
| Benchmark | GPT-5.1 High | Seed 2.1 Pro |
|---|---|---|
| MMMU-Pro | — | 82.7 |
| CharXiv Reasoning | — | 86.4 |
agent
| Benchmark | GPT-5.1 High | Seed 2.1 Pro |
|---|---|---|
| OSWorld | — | 78.8 |
| BrowseComp | — | 86.2 |
| Toolathlon | — | 50.6 |
| MCP Atlas | — | 83.8 |
| APEX Agents |
Category Winners
coding
B
Seed 2.1 Pro71.0
math
83.5
reasoning
B
Seed 2.1 Pro92.2
multimodal
B
Seed 2.1 Pro84.3
agent
B
Seed 2.1 Pro80.3