Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude 3.5 Haiku | Phi-3.5-MoE-instruct |
|---|---|---|
| GPQA Diamond | 41.6 | 36.8 |
coding
| Benchmark | Claude 3.5 Haiku | Phi-3.5-MoE-instruct |
|---|---|---|
| SWE-bench Verified | 40.6 | — |
agent
| Benchmark | Claude 3.5 Haiku | Phi-3.5-MoE-instruct |
|---|---|---|
| TAU-Bench Retail | 51 | — |
general
| Benchmark | Claude 3.5 Haiku | Phi-3.5-MoE-instruct |
|---|---|---|
| MMMLU | — | 69.9 |
Category Winners
coding
7.0
reasoning
10.4
general
8.4
agent
10.1