Qwen3 32B vs Llama 3.2 90B Vision Instruct
Qwen3 32B is cheaper on output tokens, while Qwen3 32B offers a larger context window. Choose Qwen3 32B or Llama 3.2 90B Vision Instruct based on the trade-off between cost, context, and the benchmarks that matter for your use case.
| Spec | Qwen3 32B | Llama 3.2 90B Vision Instruct |
|---|---|---|
| Provider | Qwen | Vercel AI Gateway |
| Input / 1M tokens | $0.08 | $0.72 |
| Output / 1M tokens | $0.28 | $0.72 |
| Context window | 131K | 128K |
| Parameters | — | — |
| Open weights | — | No |
| Released | — | Sep 2024 |
| LiveBench score | 42.7 | — |
| Aider Polyglot score | 40.0 | — |
| LMArena (Chatbot Arena) Elo score | 1340 | — |
| GPQA Diamond score | 65.7 | 41.0 |
| Epoch Capabilities Index score | 138 | — |
| LMArena Vision Elo score | — | 1000 |
FAQ
Is Qwen3 32B or Llama 3.2 90B Vision Instruct cheaper?
Qwen3 32B is cheaper on output tokens ($0.28 vs $0.72 per 1M).
Which has the larger context window, Qwen3 32B or Llama 3.2 90B Vision Instruct?
Qwen3 32B has the larger context window (131K tokens).
Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.