grok-4-fast-non-reasoning
- Input / 1M
- $0.20
- Output / 1M
- $0.50
- Cached input
- $0.05
- Context
- 2M
A 1,000-token prompt costs $0.00020 to send. Token counts for xAI models are estimated — approximated, because this family has no public tokenizer.
- Tool calling
- Max output 2M
Long-context rate
Past 128K tokens in a single request, grok-4-fast-non-reasoning switches to $0.40 input and $1.00 output per million — 2.0× the base input rate. Most calculators quote only the base figure.
Scheduled for deprecation on 2026-05-15. Plan a migration before then.
Priced closest to this
| Model | Provider | Input | Output | |
|---|---|---|---|---|
| gpt-5.4-nano | OpenAI | $0.20 | $1.25 | |
| gpt-5.6-luna | OpenAI | $0.20 | $1.20 | |
| jamba-1.5 | AI21 | $0.20 | $0.40 | |
| jamba-1.5-mini | AI21 | $0.20 | $0.40 | |
| jamba-mini-1.6 | AI21 | $0.20 | $0.40 | |
| jamba-mini-1.7 | AI21 | $0.20 | $0.40 |