Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
language
| Benchmark | Claude 3 Opus | MiMo-V2.5-Pro |
|---|---|---|
| WinoGrande | 85.4 | — |
| HellaSwag | 95.4 | — |
safety
| Benchmark | Claude 3 Opus | MiMo-V2.5-Pro |
|---|---|---|
| TruthfulQA | 72.8 | — |
reasoning
| Benchmark | Claude 3 Opus | MiMo-V2.5-Pro |
|---|---|---|
| ARC-Challenge | 96.4 | — |
| GPQA Diamond | 50.4 | 66.7 |
| BigBench-Hard | 86.8 | — |
| Humanity's Last Exam | — | 34 |
general
| Benchmark | Claude 3 Opus | MiMo-V2.5-Pro |
|---|---|---|
| IFEval | 85.2 | — |
| MMLU-Pro | 68.5 | — |
| Chatbot Arena Elo | 1232 | — |
coding
| Benchmark | Claude 3 Opus | MiMo-V2.5-Pro |
|---|---|---|
| HumanEval | 84.9 | — |
| SWE-bench Verified | — | 78.9 |
| SWE-Bench Pro | — | 57.2 |
Category Winners
coding
X
MiMo-V2.5-Pro63.0
math
44.3
reasoning
51.1
general
56.6
language
75.0
safety
92.9