Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Magistral Small 2506 | Sarvam-105B |
|---|---|---|
| GPQA Diamond | 68.2 | 78.7 |
| Humanity's Last Exam | — | 11.2 |
math
| Benchmark | Magistral Small 2506 | Sarvam-105B |
|---|---|---|
| AIME 2025 | 62.8 | 96.7 |
coding
| Benchmark | Magistral Small 2506 | Sarvam-105B |
|---|---|---|
| SWE-bench Verified | — | 45 |
agent
| Benchmark | Magistral Small 2506 | Sarvam-105B |
|---|---|---|
| BrowseComp | — | 49.5 |
Category Winners
coding
S
Sarvam-105B11.3
math
S
Sarvam-105B79.9
reasoning
53.0
agent
S
Sarvam-105B22.8