Meta: Llama 3.1 70B Instruct wins on more metrics (6 of 8), but the right pick depends on what you optimize for — see the breakdown below.
| Metric | AllenAI: Olmo 3 32B Think | Meta: Llama 3.1 70B Instruct |
|---|---|---|
| Intelligence Index | 12.1 | 12.5✓ |
| Coding Index | 10.5 | 10.9✓ |
| GPQA Diamond | 61%✓ | 41% |
| Design Arena Elo | — | — |
| Speed (tokens/sec) | — | 29✓ |
| Latency | — | 290ms✓ |
| Input price /M | $0.150✓ | $0.400 |
| Output price /M | $0.500 | $0.400✓ |
| Context window | 66K | 131K✓ |
| Capabilities | ReasoningJSON | ToolsJSON |