AI Rate CardGrok 4.20 (Non-Reasoning) vs Grok 4.20 (Reasoning): API price and specs
| Field | Grok 4.20 (Non-Reasoning) | Grok 4.20 (Reasoning) |
|---|---|---|
| Lowest input, per 1M | $1.00 | $1.00 |
| Median input, per 1M | $1.25 | $1.25 |
| Lowest output, per 1M | $2.00 | $2.00 |
| Median output, per 1M | $2.50 | $2.50 |
| Lowest unit price | — | — |
| Cost of 1M in + 200k out | $1.40 | $1.40 |
| Context | 2M | 2M |
| Max output | 2M | 2M |
| Providers | 10 | 10 |
| Regions | US | US |
| Accepts | image, pdf, text | image, pdf, text |
| Returns | text | text |
| Type | Text models | Text models |
| Free option | no | no |
| Cheapest provider | vercel | vercel |
Your own workload: how many tokens the model gets (input) and gives back (output) - for example, in a month. The "Cost" row of the table above is recalculated for these numbers, for each model.
Computed from current offers each time the page opens; the best value of a row is bold, the percent is how far behind the best.
Providers selling both
aihubmix, aiml, azure, edenai, google-vertex, impossibl, jiekou, trustedrouter, vercel, xai
Frequently asked questions
Which has the larger context, Grok 4.20 (Non-Reasoning) or Grok 4.20 (Reasoning)?
Grok 4.20 (Non-Reasoning): 2M tokens; Grok 4.20 (Reasoning): 2M tokens.
Where can I use both Grok 4.20 (Non-Reasoning) and Grok 4.20 (Reasoning)?
aihubmix, aiml, azure, edenai, google-vertex, impossibl, jiekou, trustedrouter, vercel, xai