Grok 4.20 Reasoning
Grok 4.20 Reasoning is xAI's reasoning model in the Grok 4.20 beta generation, supporting extended thinking for complex problem-solving with low hallucination rates and strict prompt adherence.
import { streamText } from 'ai'
const result = streamText({ model: 'xai/grok-4.20-reasoning', prompt: 'Why is the sky blue?'})Playground
Try out Grok 4.20 Reasoning by xAI. Usage is billed to your team at API rates. Free users (those who haven't made a payment) get $5 of credits every 30 days.
Providers
Route requests across multiple providers. Copy a provider slug to set your preference. Visit the docs for more info. Using a provider means you agree to their terms, listed under Legal.
P50 throughput on live AI Gateway traffic, in tokens per second (TPS). Visit the docs for more info.
P50 time to first token (TTFT) on live AI Gateway traffic, in milliseconds. View the docs for more info.
Direct request success rate on AI Gateway and per-provider. Visit the docs for more info.
More models by xAI
| Model |
|---|
| 256K | 0.2s | 173tps | $1.00/M | $2.00/M | Read: $0.2/M Write: — | $5/K + input costs | — |  | 05/20/2026 | ||||
| 1M | 0.7s | 94tps | $1.25/M | $2.50/M | Read: $0.2/M Write: — | $5/K + input costs | — | | 04/30/2026 | ||||
| 2M | 2.2s | 2178tps | $1.25/M | $2.50/M | Read: $0.2/M Write: — | $5/K + input costs | — | | 03/11/2026 | ||||
| 2M | 0.4s | 143tps | $1.25/M | $2.50/M | Read: $0.2/M Write: — | $5/K + input costs | — |  | 03/09/2026 | ||||
| 1M | 1.0s | 186tps | $0.20/M | $0.50/M | Read:$0.05/M Write:— | — | — | | 07/09/2025 | ||||
| 1M | 0.3s | 61tps | $0.20/M | $0.50/M | Read:$0.05/M Write:— | — | — | | 07/09/2025 |
About Grok 4.20 Reasoning
Grok 4.20 Reasoning was released March 9, 2026 as the reasoning-enabled variant of xAI's Grok 4.20 beta generation. It supports extended thinking for complex problem-solving tasks, generating chain-of-thought reasoning traces before producing final answers. The model maintains the low hallucination rates and strict prompt adherence from the Grok 4.20 foundation.
The reasoning capability improves accuracy on tasks that benefit from structured thinking: mathematical problem solving, code debugging, logical analysis, and multi-step planning. Reasoning traces are visible in the API response, providing transparency you can use for debugging and auditing.
This model is currently in beta.
What To Consider When Choosing a Provider
- Configuration: Grok 4.20 Reasoning is in beta. Expect potential changes before general availability.
- Configuration: Extended thinking adds tokens and latency to each response. For tasks that don't benefit from structured reasoning, the non-reasoning variant is faster and cheaper.
- Zero Data Retention: AI Gateway supports Zero Data Retention for this model via direct gateway requests (BYOK is not included). To configure this, check the documentation.
- Authentication: AI Gateway authenticates requests using an API key or OIDC token. You do not need to manage provider credentials directly.
When to Use Grok 4.20 Reasoning
Best For
- Complex problem solving: Requiring multi-step reasoning and extended thinking
- Mathematical and scientific analysis: Structured chain-of-thought improves accuracy
- Code debugging and review: That benefits from step-by-step logical analysis
- High-stakes decisions: Low hallucination rates and traceable reasoning are critical
- Research and analysis tasks: Operating over large document collections within the context of 2M tokens
Consider Alternatives When
- Speed-focused tasks: The Grok 4.20 Non-Reasoning variant delivers faster responses at lower cost
- Multi-agent orchestration: The Grok 4.20 Multi-Agent variant is purpose-built for agent collaboration
- Stable production deployments: Beta models introduce risk. Use Grok 4.1 Fast Reasoning or the full Grok 4 instead
- Simple direct-response tasks: Reasoning overhead adds cost without improving quality
Conclusion
Grok 4.20 Reasoning brings extended thinking to the Grok 4.20 beta generation. Its combination of chain-of-thought reasoning, low hallucination rates, and strict prompt adherence makes it suitable for complex analytical tasks where accuracy and transparency matter. Consider the beta status when planning production use.
Frequently Asked Questions
What does 'reasoning' mean for Grok 4.20 Reasoning?
Grok 4.20 Reasoning generates chain-of-thought reasoning traces before producing final answers. This extended thinking improves accuracy on complex tasks at the cost of additional latency and tokens.
What is the context window for Grok 4.20 Reasoning?
2M tokens.
Is Grok 4.20 Reasoning in beta?
Yes. Grok 4.20 Reasoning is currently in beta. Behavior, pricing, or availability may change before general availability.
What does Grok 4.20 Reasoning cost?
Rates are listed on this page. They reflect the providers routing through AI Gateway and shift when providers update their pricing.
How do I authenticate with Grok 4.20 Reasoning through Vercel AI Gateway?
Use your Vercel AI Gateway API key with
xai/grok-4.20-reasoningas the model identifier. AI Gateway manages routing and authentication automatically.Can I see the reasoning traces?
Yes. Chain-of-thought traces are included in the API response. You can inspect and log them for debugging and auditing.
Does Vercel AI Gateway support Zero Data Retention for Grok 4.20 Reasoning?
Yes, Zero Data Retention is available for this model. Zero Data Retention is offered on a per-provider basis. See https://vercel.com/docs/ai-gateway/capabilities/zdr for details.