THE DEEPSEEK MODEL FAMILY
DeepSeek API models & pricing
Looking for a cheap DeepSeek API? Compare AELDER's catalog versions and their token rates, including the 1017 release of DeepSeek V4.1 Flash. Use one OpenAI-compatible key and pay for the tokens you use.
- Models in this catalog
- 3
- Input from / 1M tokens
- $0.0135
- Lower token rates, up to
- 91%
Savings compare input/output unit rates with the stated official tier. Actual cost depends on your selected model and token usage.
Compare API token prices
Calculate your savingsPrices per 1M tokens. Cached tokens cost 10% of the input price. Very small requests are charged a minimum of $0.00002. Struck-through prices are official list prices. Official cached-input prices are shown only when confirmed.
Keep the DeepSeek release explicit.
This catalog lists distinct DeepSeek records. A similar name does not make the release, benchmark reference or request ID the same. Open the model page before comparing versions or configuring a client.
V4.1 Flash · 1017
The V4.1 Flash entry is the 1017 release. Its model page identifies the version so it is not confused with the earlier September release.
V4 Pro
The Pro entry has its own rates and published specifications. The Pro name alone does not establish a higher benchmark score than a later Flash release.
V4 Flash
This earlier catalog entry has a clearly labelled historical April 0420 benchmark reference. It is separate from the V4.1 Flash 1017 page.
DeepSeek API pricing: how token costs work.
The V4.1 Flash 1017 and V4 Pro official comparisons use off-peak reference rates; official peak rates are higher. These reference tiers do not define AELDER billing. Cached input has a separate rate and depends on a qualifying cache hit.
Input is the text and context you send; output is what the model generates. Cached input is a separate unit rate, not a promise that every request will hit the cache.
Compare the model and the service.
For a provider comparison, read AELDER vs OpenRouter API pricing. That page uses its own dated OpenRouter reference; the crossed-out prices here are official model rates.
Estimate a token mix with the API cost calculator, or browse all model families and API prices.
Connect to the DeepSeek API through AELDER.
Set your AELDER API key, the service Base URL and an exact model ID from your account's model list. A catalog name is not a guarantee of every optional API feature.
Use the Python and JavaScript OpenAI SDK guide for an OpenAI-compatible Chat Completions connection.
Connect a model in n8n →
Use the OpenAI credential, service Base URL and exact model ID. Start with the small text workflow before testing AI-agent tools or larger automations.
Need another tool? Browse the integration guides. Confirm model and protocol support before relying on tools, images or structured output.
DeepSeek API questions
Which DeepSeek V4.1 Flash version is listed here?
The DeepSeek V4.1 Flash record is the 1017 release. Its page shows the version and catalog ID. Confirm the exact request ID and availability in your account before making a request.
Does the off-peak label mean AELDER prices change by time of day?
The label identifies the official comparison tier used for those models. It does not describe a time-based AELDER billing rule. Our current input, cached input and output rates are the prices shown in the table.