BroadRouter

DeepSeek V3.1 vs DeepSeek V3 API pricing

Updated

DeepSeek V3.1: USD 0.25 input, USD 0.95 output per 1M tokens (DeepInfra). DeepSeek V3: USD 0.25 input, USD 1 output per 1M tokens (SiliconFlow). At 3 input tokens per output token, DeepSeek V3 costs 2.9% more than DeepSeek V3.1 (USD 0.4375 vs USD 0.425 per 1M).

DeepSeek V3 is the release before DeepSeek V3.1 in the same DeepSeek line (by version in the name), so this page shows what the newer release costs. Price only: every number is computed from prices read from each provider's own pricing page (linked per cell, newest check 2026-11-15), with no claim about quality, speed or accuracy.

How do DeepSeek V3.1 and DeepSeek V3 differ on price?

DeepSeek V3.1 and DeepSeek V3 prices side by side

USD per 1M tokens, standard tier. "Cheapest" is the lowest price at 3 input tokens to 1 output token; each line can come from a different provider.

DeepSeek V3.1 and DeepSeek V3 API prices compared
PriceDeepSeek V3.1DeepSeek V3
Model pageDeepSeek V3.1
DeepSeek · 7 providers
DeepSeek V3
DeepSeek · 3 providers
Cheapest standard price, blended 3:1USD 0.425 lower
USD 0.25 in / USD 0.95 out at DeepInfra
DeepInfra page checked 2026-11-14
USD 0.4375
USD 0.25 in / USD 1 out at SiliconFlow
SiliconFlow page checked 2026-11-14
Cheapest inputUSD 0.25 tie
DeepInfra source
USD 0.25 tie
SiliconFlow source
Cheapest cached inputUSD 0.06
Google Vertex AI source
No standard price lists a cache price; cached tokens bill at the input price.
Cheapest outputUSD 0.95
DeepInfra source
USD 0.89 lower
DeepInfra source
Maker's own APIDeepSeek publishes no paid standard price on its own API that we found.DeepSeek publishes no paid standard price on its own API that we found.
Median standard priceUSD 0.285 / USD 1
input / output, 7 prices
USD 0.32 / USD 1
input / output, 3 prices
Context window163,840 tokens, largest stated; from 131,072 at some providers163,840 tokens, largest stated
Providers with a standard price73

Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).

What do DeepSeek V3.1 and DeepSeek V3 cost for a real workload?

30 days at each model's cheapest standard price for the workload (presets of the cost calculator; cached input at the cached price where listed).

Monthly cost of three workloads on DeepSeek V3.1 and DeepSeek V3
WorkloadDeepSeek V3.1DeepSeek V3Difference
Support chatbot
2,000 requests a day, 1,500 input tokens (60% cached), 300 output
USD 33.12
DeepInfra
USD 40.50
SiliconFlow
USD 7.38
DeepSeek V3.1 lower
RAG over documents
1,000 requests a day, 8,000 input tokens (30% cached), 500 output
USD 65.61
DeepInfra
USD 75.00
SiliconFlow
USD 9.39
DeepSeek V3.1 lower
Coding agent (one developer)
1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning
USD 351.90
DeepInfra
USD 540.00
SiliconFlow
USD 188.10
DeepSeek V3.1 lower

DeepSeek V3.1 is cheaper on all three workloads. The widest gap is on Coding agent (one developer): the dearer model costs 53% more than the cheaper one.

Does the cheaper model change with your input:output ratio?

The cheaper model changes with the mix: below 0.86 input tokens per output token, DeepSeek V3 is cheaper; above 0.86, DeepSeek V3.1 is cheaper. At the site's 3:1 blend, DeepSeek V3.1 is cheaper. More input per output token, as with retrieval or agents that resend long contexts, favours the lower input price; long answers favour the lower output price.

Price per 1M tokens of DeepSeek V3.1 and DeepSeek V3 at different input:output mixes
Input : output tokensDeepSeek V3.1DeepSeek V3
1 : 3 (long answers)USD 0.775
DeepInfra
USD 0.7475 lower
DeepInfra
1 : 1USD 0.60 lower
DeepInfra
USD 0.605
DeepInfra
3 : 1 (the site’s blend)USD 0.425 lower
DeepInfra
USD 0.4375
SiliconFlow
10 : 1 (retrieval, long prompts)USD 0.3136 lower
DeepInfra
USD 0.3182
SiliconFlow
30 : 1 (agents, big contexts)USD 0.2726 lower
DeepInfra
USD 0.2742
SiliconFlow

USD per 1M tokens at the cheapest standard provider for each mix, before cache discounts. Thinking tokens bill as output.

What do DeepSeek V3.1 and DeepSeek V3 cost per character in other languages?

Not measured: the real cost by language dataset covers only tokenizers we could measure from published files, and DeepSeek V3 is not among them. Count tokens on your own text with the counters below.

Check it on your own traffic

More comparisons: DeepSeek V3.2 vs DeepSeek V3.1 · DeepSeek V3 0324 vs DeepSeek V3 · every DeepSeek V3.1 comparison · every DeepSeek V3 comparison · all 662 price comparisons

Questions

Is DeepSeek V3.1 cheaper than DeepSeek V3?

Yes. On a blend of 3 input tokens to 1 output token, DeepSeek V3.1 costs USD 0.425 per 1M tokens and DeepSeek V3 USD 0.4375, at each model's cheapest standard price. For input alone, both start at USD 0.25 per 1M; for output alone, DeepSeek V3 is (USD 0.89 against USD 0.95). This compares price only, not quality.

What do DeepSeek V3.1 and DeepSeek V3 cost per 1M tokens?

DeepSeek V3.1: from USD 0.25 input and USD 0.95 output at DeepInfra; median of 7 standard prices USD 0.285 / USD 1. DeepSeek V3: from USD 0.25 input and USD 1 output at SiliconFlow; median of 3 standard prices USD 0.32 / USD 1. Each price is read from the provider's own pricing page; the source is linked in the table.

Which providers sell both DeepSeek V3.1 and DeepSeek V3?

3 providers have a standard price for both: DeepInfra (USD 0.425 and USD 0.4625 at 3:1), SiliconFlow (USD 0.4525 and USD 0.4375 at 3:1), Microsoft Azure (USD 2.158 and USD 1.995 at 3:1). The first is the lowest combined price. Buying each model from its own cheapest provider costs USD 0.425 and USD 0.4375.

Which is cheaper for a coding agent or a chatbot, DeepSeek V3.1 or DeepSeek V3?

For a coding agent (1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning): USD 351.90 a month on DeepSeek V3.1 (DeepInfra) against USD 540.00 on DeepSeek V3 (SiliconFlow), a difference of USD 188.10. For a support chatbot (2,000 requests a day, 1,500 input tokens (60% cached), 300 output): USD 33.12 a month on DeepSeek V3.1 (DeepInfra) against USD 40.50 on DeepSeek V3 (SiliconFlow), a difference of USD 7.38. Each uses the cheapest standard price for that workload, cached input included. Try your own numbers.

At what input:output ratio does the cheaper of DeepSeek V3.1 and DeepSeek V3 change?

The cheaper model changes with the mix: below 0.86 input tokens per output token, DeepSeek V3 is cheaper; above 0.86, DeepSeek V3.1 is cheaper. At the site's 3:1 blend, DeepSeek V3.1 is cheaper. More input per output token, as with retrieval or agents that resend long contexts, favours the lower input price; long answers favour the lower output price.

Do DeepSeek V3.1 and DeepSeek V3 have cached input pricing?

DeepSeek V3.1 from USD 0.06 per 1M cached input tokens (Google Vertex AI; 3 of 7 standard prices list a discount). DeepSeek V3: no standard price lists a cached-input discount. Cached tokens are billed at the input price where no cache price is published.

Which has the larger context window, DeepSeek V3.1 or DeepSeek V3?

Both go up to 163,840 tokens on the providers that state it.

Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity