Grok 4.20 (Reasoning) pricing
xAI · API id grok-4.20-0309-reasoning · verified 2026-10-18
| Input | $1.25 |
| Cached input (cache hit) | $0.20 |
| Cache write | — |
| Output | $2.50 |
| Batch discount | 20% |
| Context window | 1,000,000 tokens |
| Max output | — |
| Above 200K prompt tokens | $2.50 in / $5.00 out |
USD per 1M tokens unless noted. Official source
Estimate your workload on Grok 4.20 (Reasoning) →
What Grok 4.20 (Reasoning) costs for real workloads
| Workload | Grok 4.20 (Reasoning) |
|---|---|
| 10,000 chatbot replies 2,000 input + 400 output tokens each, 50% of the prompt cached | $24.50 |
| One 30-step coding agent run 12K system + tools prompt, 1K output and 3K tool result per step, caching on | $0.64 |
| 1M documents classified (batch) 600 input + 20 output tokens each, batch API where available | $640 |
Compared with the cheapest other reasoning models. Change the numbers in the calculator.
Billing notes
Non-reasoning variant grok-4.20-0309-non-reasoning has identical prices. Batch = 20% off. Requests with >=200K prompt tokens bill ALL tokens at the higher rate. Grok 4.7 Fast is not on the public API.
Where Grok 4.20 (Reasoning) ranks by use case
Position by cost among all models that fit each workload (1 = cheapest).
- Cheapest LLM for coding agents: #18 of 36 (Per run (cached): $0.64)
- Cheapest LLM for RAG chatbots: #17 of 36 (Per 1,000 answers: $6.92)
- Cheapest LLM for batch classification and extraction: #17 of 36 (Classify 1M docs: $640)
- Long-context document analysis (400K tokens): #15 of 29 (10 questions, cached: $2.55)
- Cheapest LLM for high-volume customer support: #16 of 36 (Per 1,000 replies: $2.30)
Counting tokens for a prompt? The token counter estimates Grok 4.20 (Reasoning) tokens (xAI, DeepSeek, Mistral tokenizer, estimated) and input cost.
Other xAI models
- Grok 4.7: $2.00 / $6.00
- Grok 4.3: $1.25 / $2.50
- Grok Build 0.1: $1.00 / $2.00
FAQ
How much does Grok 4.20 (Reasoning) cost per 1M tokens?
Grok 4.20 (Reasoning) costs $1.25 per 1M input tokens and $2.50 per 1M output tokens at standard rates (verified 2026-10-18).
Does Grok 4.20 (Reasoning) support prompt caching?
Yes. Cached input is billed at $0.20 per 1M tokens.
Is there a batch discount for Grok 4.20 (Reasoning)?
Yes. The batch API is 20% cheaper than standard requests.
What is the context window of Grok 4.20 (Reasoning)?
1,000,000 tokens. Prompts over 200,000 tokens are billed at $2.50 input and $5.00 output for the whole request.