# Grok 4.20 (non-reasoning) vs Grok 4.7

Grok 4.7 is one of xAI's newest models and Grok 4.20 (non-reasoning) an earlier one, so this is also what moving up costs or saves. Grok 4.20 (non-reasoning) is the cheaper of the 2: for the same tokens, 3 read for each 1 written, Grok 4.7 costs 1.9 times as much. Below are both price lists, what 8 everyday jobs cost a month on each, and how many tokens the same text takes on each, which can change the answer.

- Page, with a calculator for your own job: https://tecyfy.com/prices/compare/grok-4-20-0309-non-reasoning-vs-grok-4-7
- Makers: Grok 4.20 (non-reasoning) from xAI, Grok 4.7 from xAI

## Price per million tokens, in US dollars

| | Grok 4.20 (non-reasoning) | Grok 4.7 |
| --- | --: | --: |
| Input | $1.25 | $2 |
| Cached input | $0.20 | $0.50 |
| Cache write | not listed | not listed |
| Output | $2.50 | $6 |
| Blend, 3 read for each 1 written | $1.5625 | $3 |
| Batch input | not listed | not listed |
| Batch output | not listed | not listed |
| Long prompt input (above 200K on Grok 4.20 (non-reasoning), 200K on Grok 4.7) | $2.50 | $4 |
| Long prompt output | $5 | $12 |
| Context window, tokens | 1M | 500K |
| Capability score (Epoch Capabilities Index; GPT-5 is 150) | not listed | 154 |
| Longest answer, tokens | not listed | not listed |

## What 8 everyday jobs cost a month on each

Each job at its usual runs a day, for 30 days, at the standard price.

| Job | Tokens in, out, runs a day | Grok 4.20 (non-reasoning) | Grok 4.7 |
| --- | --- | --: | --: |
| A chat question | 300, 500, 100 | $4.88 | $10.80 |
| A hard question, with thinking | 300, 1,400, 100 | $11.63 | $27.00 |
| A long report, summarised | 60,000, 2,000, 20 | $48.00 | $79.20 |
| A 40-turn agent task | 2,800,000, 40,000, 10 | $286 | $618 |
| A support bot's reply | 4,000, 300, 2,000 | $219 | $408 |
| Sort a message into a category | 250, 5, 50,000 | $488 | $795 |
| Read an invoice into fields | 1,500, 250, 1,000 | $75.00 | $135 |
| Review a pull request | 25,000, 4,500, 50 | $63.75 | $116 |

## Questions

### Which is cheaper, Grok 4.20 (non-reasoning) or Grok 4.7?

Grok 4.20 (non-reasoning) is cheaper: $1.25 per million tokens read and $2.50 per million written, against $2 and $6 on Grok 4.7. On a mix of 3 tokens read for each 1 written, Grok 4.7 costs 1.9 times as much.

### Which is better, Grok 4.20 (non-reasoning) or Grok 4.7?

Grok 4.7 scores 154 on the Epoch Capabilities Index, an independent score from Epoch AI that combines over 50 benchmarks (GPT-5 is 150). Grok 4.20 (non-reasoning) has no independent score yet, so the only fair test is to run the same 10 tasks on both.

### What does a support bot cost a month on Grok 4.20 (non-reasoning) and Grok 4.7?

At 2,000 replies a day, each reading 4,000 tokens (50 percent of them from cache) and writing 300, it costs $219 a month on Grok 4.20 (non-reasoning) and $408 on Grok 4.7: Grok 4.7 costs 1.9 times as much for this job.

### What does an AI agent task cost on Grok 4.20 (non-reasoning) and Grok 4.7?

One 40-turn task that reads 2,800,000 tokens in all, 90 percent of them from cache, and writes 40,000, costs 95 cents on Grok 4.20 (non-reasoning) and $2.06 on Grok 4.7. At 10 tasks a day that is $286 and $618 a month.

### Which has the larger context window, Grok 4.20 (non-reasoning) or Grok 4.7?

Grok 4.20 (non-reasoning), with 1M tokens against 500K on Grok 4.7. The window holds your instructions, the files you attach, the conversation so far and the answer, all at once.

### Does the same text take the same number of tokens on Grok 4.20 (non-reasoning) and Grok 4.7?

Yes, both are counted the same way: our lesson "How AI writes", 1,700 words long, is about 2,100 tokens on each.

### Is it cheaper to move from Grok 4.20 (non-reasoning) to Grok 4.7?

No, it costs more per token: Grok 4.7 charges $2 per million tokens read and $6 written, against $1.25 and $2.50 on Grok 4.20 (non-reasoning), 1.9 times as much for the same tokens. It pays off only if better answers save more than that, for example in fewer retries.

### Can I pay less on Grok 4.20 (non-reasoning) and Grok 4.7 for work that can wait?

Not with a batch price: neither maker's price page lists one for these models. A cache still lowers the price of text the model reads again.

## Sources

- [xAI: API prices](https://docs.x.ai/docs/models)
- [Epoch Capabilities Index, by Epoch AI (CC BY 4.0)](https://epoch.ai/benchmarks/eci)

## Other comparisons

- [Grok 4.6 vs Grok 4.7](https://tecyfy.com/prices/compare/grok-4-6-vs-grok-4-7.md)
- [Grok 4.5 vs Grok 4.7](https://tecyfy.com/prices/compare/grok-4-5-vs-grok-4-7.md)
- [Grok 4.3 vs Grok 4.7](https://tecyfy.com/prices/compare/grok-4-3-vs-grok-4-7.md)
- [Grok 4.20 (reasoning) vs Grok 4.7](https://tecyfy.com/prices/compare/grok-4-20-0309-reasoning-vs-grok-4-7.md)
- [Grok 4.7 vs Grok Build 0.1](https://tecyfy.com/prices/compare/grok-4-7-vs-grok-build-0-1.md)
- [Grok 4.20 Multi-agent vs Grok 4.7](https://tecyfy.com/prices/compare/grok-4-20-multi-agent-0309-vs-grok-4-7.md)