DeepSeek V4 Pro vs DeepSeek V4 Flash API pricing
Updated
DeepSeek V4 Pro: USD 0.50 input, USD 2 output per 1M tokens (Relace). DeepSeek V4 Flash: USD 0.07 input, USD 0.14 output per 1M tokens (Relace). At 3 input tokens per output token, DeepSeek V4 Pro costs 10× as much as DeepSeek V4 Flash (USD 0.875 vs USD 0.0875 per 1M).
How do DeepSeek V4 Pro and DeepSeek V4 Flash differ on price?
- Cheapest input price: DeepSeek V4 Flash USD 0.07 against USD 0.50 per 1M tokens, so DeepSeek V4 Pro costs 7.1× as much.
- Cheapest output price: DeepSeek V4 Flash USD 0.14 against USD 1.914 per 1M tokens, so DeepSeek V4 Pro costs 14× as much.
- Among the 201 indexed models, DeepSeek V4 Pro ranks 124th cheapest and DeepSeek V4 Flash 14th on the 3:1 blend.
- No input:output mix changes the answer: DeepSeek V4 Flash is cheaper at every mix before cache discounts.
- Both list cached-input discounts: DeepSeek V4 Pro at 11 standard prices of 12, from USD 0.05 (Relace, 90% below that provider's input price); DeepSeek V4 Flash at 10 standard prices of 10, from USD 0.007 (Reka, 94% below that provider's input price).
- Both state a context window of up to 1,048,576 tokens.
- DeepSeek V4 Pro has 12 providers with a standard price, 2 more than DeepSeek V4 Flash; 9 providers sell both (Relace, GMI Cloud, DeepInfra, SiliconFlow and 5 more).
- On one account, Relace charges USD 0.875 for DeepSeek V4 Pro (its cheapest price) and USD 0.0875 for DeepSeek V4 Flash (its cheapest price) at 3:1, the lowest combined price among the providers that sell both.
- Shopping around matters more for DeepSeek V4 Pro: its dearest standard price is 3.4× its cheapest, against 3.1× for DeepSeek V4 Flash.
- For the one-developer coding agent preset, the gap is USD 4,063 a year in DeepSeek V4 Flash's favour (DeepSeek V4 Pro USD 391.50 a month, DeepSeek V4 Flash USD 52.92).
- Per character of input, DeepSeek V4 Pro costs 7.1× as much as DeepSeek V4 Flash in every measured language, because both use the DeepSeek V4.1 Flash tokenizer.
DeepSeek V4 Pro and DeepSeek V4 Flash 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.
| Price | DeepSeek V4 Pro | DeepSeek V4 Flash |
|---|---|---|
| Model page | DeepSeek V4 Pro DeepSeek · 12 providers | DeepSeek V4 Flash DeepSeek · 10 providers |
| Cheapest standard price, blended 3:1 | USD 0.875 USD 0.50 in / USD 2 out at Relace Relace page checked 2026-11-16 | USD 0.0875 lower USD 0.07 in / USD 0.14 out at Relace Relace page checked 2026-11-16 |
| Cheapest input | USD 0.50 Relace source | USD 0.07 lower Relace source |
| Cheapest cached input | USD 0.05 Relace source | USD 0.007 lower Reka source |
| Cheapest output | USD 1.914 promo GMI Cloud source | USD 0.14 lower Relace source |
| Maker's own API | USD 1.32 / USD 3.96 DeepSeek, serving version source | DeepSeek publishes no paid standard price on its own API that we found. |
| Median standard price | USD 1.625 / USD 3.2505 input / output, 12 prices | USD 0.134 / USD 0.28 input / output, 10 prices |
| Context window | 1,048,576 tokens, largest stated; from 262,144 at some providers | 1,048,576 tokens, largest stated; from 262,144 at some providers |
| Providers with a standard price | 12 | 10 |
- Lower marks the lower price on a line; tie, equal prices. Other tiers and free rows are on the model pages.
- Promo: the provider calls the price promotional or discounted, so it is current now and can end without notice.
Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).
What do DeepSeek V4 Pro and DeepSeek V4 Flash 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).
| Workload | DeepSeek V4 Pro | DeepSeek V4 Flash | Difference |
|---|---|---|---|
| Support chatbot 2,000 requests a day, 1,500 input tokens (60% cached), 300 output | USD 56.70 Relace | USD 5.80 Relace | USD 50.90 DeepSeek V4 Flash lower |
| RAG over documents 1,000 requests a day, 8,000 input tokens (30% cached), 500 output | USD 117.60 Relace | USD 14.87 Relace | USD 102.73 DeepSeek V4 Flash lower |
| Coding agent (one developer) 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning | USD 391.50 Relace | USD 52.92 Relace | USD 338.58 DeepSeek V4 Flash lower |
DeepSeek V4 Flash is cheaper on all three workloads. The widest gap is on Support chatbot: the dearer model costs 9.8× as much as the cheaper one.
Does the cheaper model change with your input:output ratio?
DeepSeek V4 Flash is cheaper at every mix of input and output tokens, before cache discounts: its cheapest input price and its cheapest output price are both at or below DeepSeek V4 Pro's, so the input:output ratio never changes the answer.
| Input : output tokens | DeepSeek V4 Pro | DeepSeek V4 Flash |
|---|---|---|
| 1 : 3 (long answers) | USD 1.625 Relace | USD 0.1225 lower Relace |
| 1 : 1 | USD 1.25 Relace | USD 0.105 lower Relace |
| 3 : 1 (the site’s blend) | USD 0.875 Relace | USD 0.0875 lower Relace |
| 10 : 1 (retrieval, long prompts) | USD 0.6364 Relace | USD 0.07636 lower Relace |
| 30 : 1 (agents, big contexts) | USD 0.5484 Relace | USD 0.07226 lower Relace |
USD per 1M tokens at the cheapest standard provider for each mix, before cache discounts. Thinking tokens bill as output.
What do DeepSeek V4 Pro and DeepSeek V4 Flash cost per character in other languages?
Cost to send 1M characters at each model's cheapest non-promotional standard input price, with the DeepSeek V4.1 Flash tokenizer for DeepSeek V4 Pro (the same one for both, so the gap equals the price gap):
| Content | DeepSeek V4 Pro | DeepSeek V4 Flash |
|---|---|---|
| English | USD 0.107 Relace | USD 0.015 lower Relace |
| Arabic MSA | USD 0.188 Relace | USD 0.026 lower Relace |
| Egyptian Arabic | USD 0.193 Relace | USD 0.027 lower Relace |
| Hindi | USD 0.289 Relace | USD 0.040 lower Relace |
| Chinese | USD 0.293 Relace | USD 0.041 lower Relace |
| Spanish | USD 0.129 Relace | USD 0.018 lower Relace |
| Code | USD 0.132 Relace | USD 0.018 lower Relace |
From the real cost by language dataset (measured tokens per character on FLORES-200 and other texts). Your own text will differ.
Published results on the same benchmark
BroadNet did not run these benchmarks. BroadRouter is BroadNet's site. Each row pairs results from one source with the same setting for both models, as published; results from different sources or settings are never put side by side. One benchmark says little about your own workload.
| Benchmark | DeepSeek V4 Pro | DeepSeek V4 Flash | Source and setting |
|---|---|---|---|
| MMLU-Pro (EM) | 87.5 | 86.2 | DeepSeek model card: DeepSeek V4 Pro, DeepSeek V4 Flash, checked 2026-11-16 Think Max reasoning mode |
| GPQA Diamond (Pass@1) | 90.1 | 88.1 | DeepSeek model card: DeepSeek V4 Pro, DeepSeek V4 Flash, checked 2026-11-16 Think Max reasoning mode |
| LiveCodeBench (Pass@1) | 93.5 | 91.6 | DeepSeek model card: DeepSeek V4 Pro, DeepSeek V4 Flash, checked 2026-11-16 Think Max reasoning mode |
| SWE Verified (Resolved) | 80.6 | 79.0 | DeepSeek model card: DeepSeek V4 Pro, DeepSeek V4 Flash, checked 2026-11-16 Think Max reasoning mode |
| Terminal Bench 2.0 (Acc) | 67.9 | 56.9 | DeepSeek model card: DeepSeek V4 Pro, DeepSeek V4 Flash, checked 2026-11-16 Think Max reasoning mode |
All published results for each model, with every setting: DeepSeek V4 Pro · DeepSeek V4 Flash.
Check it on your own traffic
- DeepSeek V4 Pro API pricing: all 12 providers and other tiers.
- DeepSeek V4 Flash API pricing: all 10 providers and other tiers.
- Cost calculator with DeepSeek V4 Pro and DeepSeek V4 Flash preloaded
- DeepSeek token counter
- Prompt caching calculator
More comparisons: DeepSeek V4 Pro vs DeepSeek V3.2 · DeepSeek V4 Pro vs GLM 5.3 · DeepSeek V4 Pro vs Qwen3.8 Max · DeepSeek V4 Pro vs Kimi K3 · GPT-6 Sol vs DeepSeek V4 Pro · DeepSeek V4 Pro 0813 vs DeepSeek V4 Pro · DeepSeek V4.1 Flash vs DeepSeek V4 Flash · Gemma 4 26B A4B vs DeepSeek V4 Flash · DeepSeek V4 Flash 0731 vs DeepSeek V4 Flash · DeepSeek V4 Flash vs DeepSeek V3.2 · every DeepSeek V4 Pro comparison · every DeepSeek V4 Flash comparison · all 662 price comparisons
Questions
Is DeepSeek V4 Pro cheaper than DeepSeek V4 Flash?
No. On a blend of 3 input tokens to 1 output token, DeepSeek V4 Flash costs USD 0.0875 per 1M tokens and DeepSeek V4 Pro USD 0.875, at each model's cheapest standard price. For input alone, DeepSeek V4 Flash is cheaper (USD 0.07 against USD 0.50 per 1M); for output alone, DeepSeek V4 Flash is (USD 0.14 against USD 1.914). This compares price only, not quality.
What do DeepSeek V4 Pro and DeepSeek V4 Flash cost per 1M tokens?
DeepSeek V4 Pro: from USD 0.50 input and USD 2 output at Relace; DeepSeek's own API charges USD 1.32 / USD 3.96; median of 12 standard prices USD 1.625 / USD 3.2505. DeepSeek V4 Flash: from USD 0.07 input and USD 0.14 output at Relace; median of 10 standard prices USD 0.134 / USD 0.28. Each price is read from the provider's own pricing page; the source is linked in the table.
Which providers sell both DeepSeek V4 Pro and DeepSeek V4 Flash?
9 providers have a standard price for both: Relace (USD 0.875 and USD 0.0875 at 3:1), GMI Cloud (USD 1.196 and USD 0.1138 at 3:1), DeepInfra (USD 1.625 and USD 0.1125 at 3:1), SiliconFlow (USD 1.91 and USD 0.1675 at 3:1), Featherless (USD 2 and USD 0.175 at 3:1), Venice (USD 2.063 and USD 0.1723 at 3:1), Baidu (USD 2.112 and USD 0.175 at 3:1), Parasail (USD 2.175 and USD 0.175 at 3:1), Microsoft Azure (USD 2.175 and USD 0.27 at 3:1). The first is the lowest combined price. Buying each model from its own cheapest provider costs USD 0.875 and USD 0.0875.
Which is cheaper for a coding agent or a chatbot, DeepSeek V4 Pro or DeepSeek V4 Flash?
For a coding agent (1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning): USD 391.50 a month on DeepSeek V4 Pro (Relace) against USD 52.92 on DeepSeek V4 Flash (Relace), a difference of USD 338.58. For a support chatbot (2,000 requests a day, 1,500 input tokens (60% cached), 300 output): USD 56.70 a month on DeepSeek V4 Pro (Relace) against USD 5.80 on DeepSeek V4 Flash (Relace), a difference of USD 50.90. 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 V4 Pro and DeepSeek V4 Flash change?
DeepSeek V4 Flash is cheaper at every mix of input and output tokens, before cache discounts: its cheapest input price and its cheapest output price are both at or below DeepSeek V4 Pro's, so the input:output ratio never changes the answer.
Do DeepSeek V4 Pro and DeepSeek V4 Flash have cached input pricing?
DeepSeek V4 Pro from USD 0.05 per 1M cached input tokens (Relace; 11 of 12 standard prices list a discount). DeepSeek V4 Flash from USD 0.007 per 1M cached input tokens (Reka; 10 of 10 standard prices list a discount). Cached tokens are billed at the input price where no cache price is published.
Which has the larger context window, DeepSeek V4 Pro or DeepSeek V4 Flash?
Both go up to 1,048,576 tokens on the providers that state it.
Which scores higher on MMLU-Pro (EM), DeepSeek V4 Pro or DeepSeek V4 Flash?
Per the DeepSeek model cards (DeepSeek V4 Pro, DeepSeek V4 Flash) (checked 2026-11-16), DeepSeek V4 Pro scores 87.5 and DeepSeek V4 Flash scores 86.2 on MMLU-Pro (EM) (Think Max reasoning mode). BroadNet did not run these benchmarks, and one benchmark says little about your own workload. Every result both models share, with sources.
Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity
Sources
- Relace: pricing (checked 2026-11-16)
- GMI Cloud: pricing (checked 2026-11-15)
- DeepSeek: own API pricing (checked 2026-11-14)
- Reka: pricing (checked 2026-11-16)
- GMI Cloud: pricing (checked 2026-11-15)
- DeepInfra: pricing (checked 2026-11-14)
- DeepInfra: pricing (checked 2026-11-14)
- SiliconFlow: pricing (checked 2026-11-14)
- Featherless: pricing (checked 2026-11-15)
- Featherless: pricing (checked 2026-11-15)
- Venice: pricing (checked 2026-11-15)
- Baidu: pricing (checked 2026-11-15)
- Parasail: pricing (checked 2026-11-14)
- Microsoft Azure: pricing (checked 2026-11-14)
- DeepSeek model card, deepseek-v4-pro (checked 2026-11-16)
- DeepSeek model card, deepseek-v4-flash (checked 2026-11-16)