← All comparisons
Qwen3.5 122B A10B (Non-reasoning) vs Llama 3.2 Instruct 1B
Alibaba vs Meta — side-by-side benchmark comparison
| Qwen3.5 122B A10B (Non-reasoning) | Llama 3.2 Instruct 1B | |
|---|---|---|
| Intelligence Index | 35.9 | 6.3 |
| Coding Index | 31.6 | 0.6 |
| Math Index | — | 0.0 |
| Output speed (tok/s) | 165.9 | 89.7 |
| Blended price ($/1M) | $1.10 | $0.05 |
| Time to first token (s) | 1.06s | 0.59s |
| aime | — | 0.0% |
| aime 25 | — | 0.0% |
| artificial analysis coding index | 31.60 | 60.0% |
| artificial analysis intelligence index | 35.90 | 6.30 |
| artificial analysis math index | — | 0.0% |
| gpqa | 82.7% | 19.6% |
| hle | 14.8% | 5.3% |
| ifbench | 50.8% | 22.8% |
| lcr | 56.0% | 5.0% |
| livecodebench | — | 1.9% |
| math 500 | — | 14.0% |
| mmlu pro | — | 20.0% |
| scicode | 35.6% | 1.7% |
| tau2 | 84.5% | 0.0% |
| terminalbench hard | 29.5% | 0.0% |
Benchmark data from Artificial Analysis.