Grok 4.20 Non-Reasoning vs o3-deep-research
Grok 4.20 Non-Reasoning is cheaper on output tokens, while Grok 4.20 Non-Reasoning offers a larger context window. Choose Grok 4.20 Non-Reasoning or o3-deep-research based on the trade-off between cost, context, and the benchmarks that matter for your use case.
| Spec | Grok 4.20 Non-Reasoning | o3-deep-research |
|---|---|---|
| Provider | Vercel AI Gateway | Vercel AI Gateway |
| Input / 1M tokens | $1.25 | $10.00 |
| Output / 1M tokens | $2.50 | $40.00 |
| Context window | 2M | 200K |
| Parameters | — | — |
| Open weights | No | No |
| Released | Mar 2026 | Jun 2025 |
FAQ
Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.