DeepSeek V3.2 vs Hy3 API pricing
Updated
DeepSeek V3.2: USD 0.2088 input, USD 0.3096 output per 1M tokens (GMI Cloud, promotional). Hy3: USD 0.132 input, USD 0.528 output per 1M tokens (SiliconFlow). At 3 input tokens per output token, DeepSeek V3.2 costs 1.3% more than Hy3 (USD 0.234 vs USD 0.231 per 1M).
How do DeepSeek V3.2 and Hy3 differ on price?
- Cheapest input price: Hy3 USD 0.132 against USD 0.2088 per 1M tokens, so DeepSeek V3.2 costs 1.58× as much (58% more).
- Cheapest output price: DeepSeek V3.2 USD 0.3096 against USD 0.528 per 1M tokens, so Hy3 costs 1.71× as much (71% more).
- Among the 201 indexed models, DeepSeek V3.2 ranks 63rd cheapest and Hy3 62nd on the 3:1 blend.
- The cheaper model changes at 2.8 input tokens per output token.
- Both list cached-input discounts: DeepSeek V3.2 at 9 standard prices of 11, from USD 0.0216 (GMI Cloud, 90% below that provider's input price); Hy3 at 3 standard prices of 4, from USD 0.033 (SiliconFlow, 75% below that provider's input price).
- Hy3 states up to 262,144 tokens of context, 1.6× the 163,840 of DeepSeek V3.2.
- DeepSeek V3.2 has 11 providers with a standard price, 7 more than Hy3; 3 providers sell both (GMI Cloud, SiliconFlow and Featherless).
- On one account, GMI Cloud charges USD 0.234 for DeepSeek V3.2 (its cheapest price) and USD 0.25 for Hy3 (1.08× its cheapest price) at 3:1, the lowest combined price among the providers that sell both.
- Shopping around matters more for Hy3: its dearest standard price is 7.6× its cheapest, against 4.3× for DeepSeek V3.2.
- DeepSeek V3.2 has a batch price at 1 provider, from USD 0.42 at 3:1 (Google Vertex AI); Hy3 has no batch price on file.
- For the one-developer coding agent preset, the gap is USD 196.34 a year in DeepSeek V3.2's favour (DeepSeek V3.2 USD 117.29 a month, Hy3 USD 133.65).
DeepSeek V3.2 and Hy3 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 V3.2 | Hy3 |
|---|---|---|
| Model page | DeepSeek V3.2 DeepSeek · 11 providers | Hy3 Tencent · 4 providers |
| Cheapest standard price, blended 3:1 | USD 0.234 promo USD 0.2088 in / USD 0.3096 out at GMI Cloud GMI Cloud page checked 2026-11-15 | USD 0.231 lower USD 0.132 in / USD 0.528 out at SiliconFlow SiliconFlow page checked 2026-11-14 |
| Cheapest input | USD 0.2088 promo GMI Cloud source | USD 0.132 lower SiliconFlow source |
| Cheapest cached input | USD 0.0216 lower promo GMI Cloud source | USD 0.033 SiliconFlow source |
| Cheapest output | USD 0.3096 lower promo GMI Cloud source | USD 0.528 SiliconFlow source |
| Cheapest non-promotional price | USD 0.26 / USD 0.38 DeepInfra, USD 0.29 blended source | The cheapest price is not promotional. |
| Maker's own API | DeepSeek publishes no paid standard price on its own API that we found. | Tencent publishes no paid standard price on its own API that we found. |
| Median standard price | USD 0.33 / USD 0.48 input / output, 11 prices | USD 0.14 / USD 0.58 input / output, 4 prices |
| Context window | 163,840 tokens, largest stated; from 131,072 at some providers | 262,144 tokens, largest stated; from 32,768 at some providers |
| Providers with a standard price | 11 | 4 |
- 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 V3.2 and Hy3 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 V3.2 | Hy3 | Difference |
|---|---|---|---|
| Support chatbot 2,000 requests a day, 1,500 input tokens (60% cached), 300 output | USD 14.26 GMI Cloud | USD 16.04 SiliconFlow | USD 1.78 DeepSeek V3.2 lower |
| RAG over documents 1,000 requests a day, 8,000 input tokens (30% cached), 500 output | USD 41.28 GMI Cloud | USD 32.47 SiliconFlow | USD 8.81 Hy3 lower |
| Coding agent (one developer) 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning | USD 117.29 GMI Cloud | USD 133.65 SiliconFlow | USD 16.36 DeepSeek V3.2 lower |
The cheaper model depends on the workload: DeepSeek V3.2 for Support chatbot, Hy3 for RAG over documents, DeepSeek V3.2 for Coding agent (one developer). Cache-heavy work rewards a low cached-input price. The widest gap is on RAG over documents: the dearer model costs 27% more than the cheaper one.
Does the cheaper model change with your input:output ratio?
The cheaper model changes with the mix: below 2.8 input tokens per output token, DeepSeek V3.2 is cheaper; above 2.8, Hy3 is cheaper. At the site's 3:1 blend, Hy3 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.
| Input : output tokens | DeepSeek V3.2 | Hy3 |
|---|---|---|
| 1 : 3 (long answers) | USD 0.2844 lower GMI Cloud | USD 0.429 SiliconFlow |
| 1 : 1 | USD 0.2592 lower GMI Cloud | USD 0.33 SiliconFlow |
| 3 : 1 (the site’s blend) | USD 0.234 GMI Cloud | USD 0.231 lower SiliconFlow |
| 10 : 1 (retrieval, long prompts) | USD 0.218 GMI Cloud | USD 0.168 lower SiliconFlow |
| 30 : 1 (agents, big contexts) | USD 0.2121 GMI Cloud | USD 0.1448 lower SiliconFlow |
USD per 1M tokens at the cheapest standard provider for each mix, before cache discounts. Thinking tokens bill as output.
What do DeepSeek V3.2 and Hy3 cost per character in other languages?
Not measured: the real cost by language dataset covers only tokenizers we could measure from published files, and Hy3 is not among them. Count tokens on your own text with the counters below.
Check it on your own traffic
- DeepSeek V3.2 API pricing: all 11 providers and other tiers.
- Hy3 API pricing: all 4 providers and other tiers.
- Cost calculator with DeepSeek V3.2 and Hy3 preloaded
- DeepSeek token counter
- Token counter
- Prompt caching calculator
More comparisons: DeepSeek V4 Pro vs DeepSeek V3.2 · Qwen3.8 27B vs DeepSeek V3.2 · DeepSeek V3.2 vs DeepSeek V3.1 · DeepSeek-V3.2-Exp vs DeepSeek V3.2 · DeepSeek V4 Flash vs DeepSeek V3.2 · Hy4 vs Hy3 · every DeepSeek V3.2 comparison · every Hy3 comparison · all 662 price comparisons
Questions
Is DeepSeek V3.2 cheaper than Hy3?
No. On a blend of 3 input tokens to 1 output token, Hy3 costs USD 0.231 per 1M tokens and DeepSeek V3.2 USD 0.234, at each model's cheapest standard price. For input alone, Hy3 is cheaper (USD 0.132 against USD 0.2088 per 1M); for output alone, DeepSeek V3.2 is (USD 0.3096 against USD 0.528). This compares price only, not quality.
What do DeepSeek V3.2 and Hy3 cost per 1M tokens?
DeepSeek V3.2: from USD 0.2088 input and USD 0.3096 output at GMI Cloud (promotional); median of 11 standard prices USD 0.33 / USD 0.48. Hy3: from USD 0.132 input and USD 0.528 output at SiliconFlow; median of 4 standard prices USD 0.14 / USD 0.58. Each price is read from the provider's own pricing page; the source is linked in the table.
Which providers sell both DeepSeek V3.2 and Hy3?
3 providers have a standard price for both: GMI Cloud (USD 0.234 and USD 0.25 at 3:1), SiliconFlow (USD 0.3075 and USD 0.231 at 3:1), Featherless (USD 0.3005 and USD 1.75 at 3:1). The first is the lowest combined price. Buying each model from its own cheapest provider costs USD 0.234 and USD 0.231.
Which is cheaper for a coding agent or a chatbot, DeepSeek V3.2 or Hy3?
For a coding agent (1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning): USD 117.29 a month on DeepSeek V3.2 (GMI Cloud) against USD 133.65 on Hy3 (SiliconFlow), a difference of USD 16.36. For a support chatbot (2,000 requests a day, 1,500 input tokens (60% cached), 300 output): USD 14.26 a month on DeepSeek V3.2 (GMI Cloud) against USD 16.04 on Hy3 (SiliconFlow), a difference of USD 1.78. 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 V3.2 and Hy3 change?
The cheaper model changes with the mix: below 2.8 input tokens per output token, DeepSeek V3.2 is cheaper; above 2.8, Hy3 is cheaper. At the site's 3:1 blend, Hy3 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 V3.2 and Hy3 have cached input pricing?
DeepSeek V3.2 from USD 0.0216 per 1M cached input tokens (GMI Cloud; 9 of 11 standard prices list a discount). Hy3 from USD 0.033 per 1M cached input tokens (SiliconFlow; 3 of 4 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 V3.2 or Hy3?
Hy3: up to 262,144 tokens, against 163,840 for DeepSeek V3.2, as stated on providers' pricing pages. Some providers offer less than the maximum; check the model pages before you pick the cheapest row for long prompts.