> ## Content Index
> Fetch the complete content index at: https://iteracompute.com/blog/llms.txt
> Use this file to discover other available public pages before exploring further.

# DeepSeek V4.1 Flash API prices cut again: compare the costs
- URL: https://iteracompute.com/blog/deepseek-v4-1-flash-api-pricing/
- Published: 2026-10-25T04:06:28.000Z
- Updated: 2026-10-25T04:06:28.000Z
- Description: We've lowered Flash input and output prices again. Compare the new rates and total request costs with DeepSeek's official peak and off-peak pricing.
- Author: IteraCompute

We've lowered the uncached-input and output prices of DeepSeek V4.1 Flash again. IteraCompute now charges $0.14 per million uncached input tokens and $0.55 per million output tokens, for both Flash routes listed below.

For a request with 20,000 input tokens and 2,000 billable output tokens, the total is $0.00390\. Below, we show what changed and compare that workload with DeepSeek's official API rates.

Prices verified October 25, 2026\. All rates are in USD per 1 million tokens. Check the current price table and Console before sending traffic.

### What's lower this time?

Compared with IteraCompute's October 21 catalog, uncached input is now 17.6% lower and output is 21.4% lower. These changes apply to both separately listed Flash model IDs.

| Token category      | October 21 catalog | Current rate | Change      |
| ------------------- | ------------------ | ------------ | ----------- |
| Uncached input / 1M | $0.17              | $0.14        | 17.6% lower |
| Output / 1M         | $0.70              | $0.55        | 21.4% lower |

The reduction is in uncached input and output. For the 20k-input / 2k-output example, the total falls from $0.00480 at the October 21 rates to $0.00390 now, a reduction of about 18.8%. Your savings depend on your mix of uncached input and output.

### Flash and Flash 1017 remain separate choices

| IteraCompute model / API ID                                                           | Uncached input / 1M | Output / 1M |
| ------------------------------------------------------------------------------------- | ------------------- | ----------- |
| [DeepSeek V4.1 Flash](https://iteracompute.com/models/deepseek-v4-1-flash/)           | $0.14               | $0.55       |
| [DeepSeek V4.1 Flash 1017](https://iteracompute.com/models/deepseek-v4-1-flash-1017/) | $0.14               | $0.55       |

Matching prices do not make these interchangeable versions. Select the exact model ID you intend to use, and check its model page for supported parameters and limits.

### How do these rates compare with DeepSeek's official API?

As checked on October 25, 2026, DeepSeek's official price table maps deepseek-flash to DeepSeek-V4.1-Flash-1017\. The version-specific comparison below therefore uses IteraCompute's Flash 1017 route, not the unversioned Flash route.

| Flash 1017 provider / time tier     | Uncached input / 1M | Output / 1M |
| ----------------------------------- | ------------------- | ----------- |
| IteraCompute — current listed rates | $0.14               | $0.55       |
| DeepSeek official — off-peak        | $0.15               | $0.60       |
| DeepSeek official — peak            | $0.30               | $1.20       |

Official peak periods are 01:00–04:00 and 06:00–10:00 UTC, Monday–Friday, excluding Chinese public holidays. Other times are off-peak, including weekends. IteraCompute's currently listed rates have no separate time tier.

This comparison covers text-token charges only; it does not establish feature parity or compare separate tools, service limits or latency.

### The total cost: 20k input + 2k output

Total cost in USD = (total input tokens) × uncached-input rate / 1,000,000

- billable output tokens × output rate / 1,000,000

At IteraCompute's current rates: 20,000 × $0.14 / 1M + 2,000 × $0.55 / 1M = $0.00390.

Using identical token counts for Flash 1017:

| Provider / time tier         | Total Cost |
| ---------------------------- | ---------- |
| IteraCompute                 | $0.00390   |
| DeepSeek official — off-peak | $0.00420   |
| DeepSeek official — peak     | $0.00840   |

For this example, IteraCompute is about 7.1% lower than official off-peak totals. Against official peak totals, the reduction is approximately 53.6%. Different token mixes produce different results.

These are arithmetic estimates, not measured bills. They exclude separate tool, storage or other fees. Include billable reasoning in output, and count every request and retry in an agent workflow.

### Connect with the OpenAI-compatible API

Create an IteraCompute account, check the model's access and balance requirements, and create an API key in the Console. Keep it on your server.

Set the OpenAI-compatible base URL to `https://api.iteracompute.com/v1` and use your IteraCompute API key.

Choose `deepseek/deepseek-v4.1-flash-1017` for the version compared above, or explicitly select `deepseek/deepseek-v4.1-flash` for the other Flash route.

Follow the maintained Flash 1017 OpenAI SDK guide or the separate Flash setup guide. Start with one small request and inspect its reply and usage.

See the API documentation for authentication, streaming, usage and error handling.

**Sources and scope.** Current rates were checked on October 25, 2026 against IteraCompute's public Models API and DeepSeek's official Models & Pricing page. The earlier IteraCompute rates come from our October 21 catalog snapshot. Comparisons use the listed USD rates without promotions or account-specific adjustments. Settle against actual usage and current billing terms.