Grok 4.20 (Reasoning) vs GPT-5.3 Codex
Grok 4.20 (Reasoning) is cheaper on output tokens, while Grok 4.20 (Reasoning) offers a larger context window. Choose Grok 4.20 (Reasoning) or GPT-5.3 Codex based on the trade-off between cost, context, and the benchmarks that matter for your use case.
| Spec | Grok 4.20 (Reasoning) | GPT-5.3 Codex |
|---|---|---|
| Provider | LLM Gateway | LLM Gateway |
| Input / 1M tokens | $2.00 | $1.75 |
| Output / 1M tokens | $6.00 | $14.00 |
| Context window | 2M | 400K |
| Parameters | — | — |
| Open weights | No | No |
| Released | Mar 2026 | Feb 2026 |
FAQ
Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.