xAI · language model
Grok 4.20: pricing, context and benchmarks
Grok 4.20 is xAI's model released on March 31, 2026. It costs $1.25 per million input tokens and $2.5 per million output tokens, with a 2M-token context window. Epoch AI ranks it #51 of 125 rated models.
| Performance (ECI) | 24/40 | |
|---|---|---|
| Price | 10/30 | |
| Context | 15/15 | |
| Features | 11/15 |
What you need to know about Grok 4.20
xAI, based in United States, put Grok 4.20 online on March 31, 2026. It accepts text, images, and files as input and produces text. Its context window reaches 2,000,000 tokens, larger than 100 % of the models in the registry, and it can write up to 1,800,000 tokens in one go.
It has been available for about 7 months, long enough for its prices and scores to have settled.
Grok 4.20 pricing
On OpenRouter, Grok 4.20 costs $1.25 per million input tokens and $2.5 per million output tokens. Counting three tokens read for every token written, its average price is $1.563 per million, more expensive than 67 % of paid models (2 times the median of $0.76).
A concrete example: a task that sends 2,000 tokens (about 1,500 words) and receives a 500-token answer costs $0.0038. At 1,000 tasks a day, budget $113 a month. The price calculator redoes this with your own volumes.
Tokens already sent and cached are billed at $0.2 per million.
Performance and benchmarks
The Epoch Capabilities Index (ECI), which aggregates published results across many benchmarks, gives Grok 4.20 a score of 152.1 (range 149.6 to 154.5). That places it #51 of 125 rated models.
Scores published by Epoch AI on the benchmarks the site tracks:
| Benchmark | What it measures | Score | Date |
|---|---|---|---|
| Terminal-Bench | real tasks in a terminal | 57.3 % | Feb 17, 2026 |
| ARC-AGI-1 | visual abstract-reasoning puzzles | 89.5 % | Feb 17, 2026 |
| ARC-AGI-2 | harder version of the ARC puzzles | 65.1 % | Feb 17, 2026 |
For at least equal performance, Claude Haiku 5.5, GPT-6 Luna, and GPT-5.6 Luna cost at least 30% less.
Features, privacy and local use
- It supports tool calling, so it can drive an AI agent, call APIs or MCP servers.
- It can answer in an enforced JSON format (structured outputs).
- It is a reasoning model: it "thinks" before answering, which helps on hard tasks but uses more output tokens.
- Its knowledge stops at September 2025.
Its weights are not published: it can only be used through the xAI API or a broker such as OpenRouter, which means sending your data off your machines.
Other xAI models
| Model | Released | Avg. price | ECI |
|---|---|---|---|
| Grok 4.7 | Sep 21, 2026 | $3 | 154.4 |
| Grok 4.6 | Aug 12, 2026 | $3 | 156.6 |
| Grok 4.5 | Jul 8, 2026 | $3 | 154.1 |
| Grok Build 0.1 | May 20, 2026 | $1.25 | — |
| Grok 4.3 | Apr 30, 2026 | $1.56 | — |
| Grok 4.20 Multi-Agent | Mar 31, 2026 | $1.56 | — |
Verdict
Computed score: 60/100, from public data (performance, price, context, features).
Strengths
- Very large context (2M tokens): long documents or a whole codebase fit in one request.
- Agent-ready: tool calling and structured JSON output.
Avoid if…
- your data must not leave your servers;
- your tasks are mostly simple: a non-reasoning model will cost less.
Alternatives
- GLM 5.2$1.78
- Gemini 3 Flash Preview$1.13
- Gemini 3.6 Flash$1.5
- Muse Spark 1.1$2
- Mistral Large 4$2.07
- Qwen3.7 Max$2.21
Comparisons and tools
Learn more
Frequently asked questions
How much does Grok 4.20 cost?
$1.25 per million input tokens and $2.5 per million output tokens on OpenRouter (checked October 25, 2026). A task with 2,000 tokens read and 500 written costs $0.0038.
How big is Grok 4.20's context window?
2,000,000 tokens, roughly 1,500,000 words, with a maximum of 1,800,000 output tokens.
Can Grok 4.20 run locally?
No, xAI does not publish its weights: it is API-only.
Is Grok 4.20 suitable for an AI agent?
Technically yes: it supports tool calling and structured outputs. Then check its cost for your volume.
How good is Grok 4.20?
Epoch AI ranks it #51 of 125 rated models (ECI 152.1). On Terminal-Bench, it scores 57.3%.
What are the alternatives to Grok 4.20?
The closest models on price and performance are GLM 5.2, Gemini 3 Flash Preview, and Gemini 3.6 Flash.
Related articles
- Mistral AI: models, pricing and agentsOctober 25, 2026
- LLM benchmarks: how to read model rankingsOctober 24, 2026
- Running a local LLM: hardware, software, and limitationsOctober 24, 2026
- Open-source or closed models for your AI agents?October 5, 2026
Neutrality. No affiliate links. No vendor pays to be reviewed, rated or ranked. Reviews are directed by François Chollet, neutral and factual, and every data point is dated and sourced. How we score