Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| Humanity's Last Exam | 64.5 | — |
| GPQA Diamond | — | 94.6 |
coding
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| SWE-bench Verified | 95 | — |
| SWE-Bench Pro | 80 | 64.6 |
multimodal
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| MMMU-Pro | — | 83 |
agent
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| BrowseComp | — | 90.4 |
| Toolathlon | — | 58 |
math
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| FrontierMath | — | 89 |
general
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| MRCR v2 | — | 91.5 |
Category Winners
coding
98.1
math
98.2
reasoning
97.2
general
97.5
multimodal
86.6
agent
85.8