DeepSeek V3.2 API pricing
DeepSeek · Budget price tier · MIT · released 2025-10. Prices in USD per 1 million tokens. Data snapshot: 2026-11-12.
A support ticket workload of 100K requests a month costs $58.00 on DeepSeek V3.2; the next cheaper model in the Budget price tier for this workload, Qwen3 Coder 30B A3B Instruct, does it for $52.50. The price we quote for DeepSeek V3.2 has changed 7 times since 2026-08-15, most recently on 2026-11-01 when the quoted host changed (-9.0%); it is now the 67th cheapest of 160 models in the Budget price tier. Buying from GMICloud rather than at the price we quote lowers the token price by about 19%, for a quantized fp8 build. It is rated a stable release but does not meet all four criteria of the FinOps Friendly badge.
- Input: $0.26 per 1M tokens
- Output: $0.38 per 1M tokens
- Batch API (in/out): Not published
- Prompt-cache read: $0.13 per 1M tokens
- Context window: 164K tokens
- Arena ELO: 1425 as of 2026-11-08
Available on
Direct API, AWS Bedrock, OpenRouter
AWS Bedrock regions: ap-northeast-1, us-east-1, us-east-2, us-west-2
Curated by hand; check with the provider before committing.
Market position
On blended price ($0.344 per 1M tokens at 30% input, 70% output) it is cheaper than 58% of models in the Budget price tier: 67th cheapest of 160. Across every price tier it is cheaper than 78% of priced text models in the catalogue: 67th cheapest of 303. Its Arena ELO of 1425 ranks 72nd of the 160 models that carry one, tied with 2 others, above 54% of them. With an Arena score of 1425, DeepSeek V3.2 clears OptimNow's Standard working quality threshold: Standard (Arena 1420+), used for meeting summary, invoice processing and call summary. Each threshold is a distance below the best ranked model, which scores 1530 on Arena scores as of 8 Nov 2026.
Licence, size and capabilities
DeepSeek V3.2 is open source: its weights are published under MIT, an OSI-approved licence with no field-of-use limit, so it can be self-hosted as well as bought by the token. Its published size is 685B parameters. Listed capabilities: Text, Reasoning, Agents. Filling its 164K context window once costs $0.0426 in input tokens at list price.
Pricing by host
A host is a company that runs the model and sells it by the token: the model maker, a cloud platform like Amazon Bedrock, Azure or Vertex, or an independent host. One model, several hosts, and often several prices. This model is sold by 6 different hosts. The price we quote elsewhere on this page is marked. We quote one host's rate so that models stay comparable; it is not a recommendation to buy there.
Cheapest host: GMICloud at $0.2088 input / $0.3096 output per 1M tokens, 19% below the price we quote on the blended price (30% input, 70% output). It serves a quantized fp8 build, which is not the same product as the full-precision model. A support ticket workload at 100K requests a month costs $46.80 there against $58.00 at the price we quote, a difference of $11.20 a month. The dearest, DigitalOcean, charges 2.7x the cheapest.
| Host | Input $/1M | Output $/1M | Precision |
|---|---|---|---|
| GMICloud | $0.2088 | $0.3096 | fp8 |
| DeepInfra (price we quote) | $0.26 | $0.38 | fp4 |
| AtlasCloud (price we quote) | $0.26 | $0.38 | fp8 |
| Baidu | $0.28 | $0.42 | fp8 |
| SiliconFlow | $0.259 | $0.80 | fp8 |
| DigitalOcean | $0.30 | $0.96 | not declared |
List prices as published on 2026-11-12, via OpenRouter. Negotiated rates, committed-use discounts and provisioned capacity are not shown. These hosts do not all serve the model at the same numeric precision, so the cheapest is not a like-for-like substitute for the dearest. GMICloud's listing is a promotional rate: 28% off its standard rate.
Monthly cost at 100K requests, by business use case
From the after-discounts figure (prompt caching on the cached input share, batch API for async workloads, only where DeepSeek publishes those rates), which assumes you implement both, to the list price. One figure means no discount is published.
| Use case | Tokens (in/out) | Monthly (after discounts – list) |
|---|---|---|
| Support Ticket | 1,500 / 500 | $46 – $58 |
| Knowledge Q&A | 2,000 / 800 | $64 – $82 |
| Meeting Summary | 10,000 / 1,200 | $293 – $306 |
| Marketing Content | 2,500 / 1,800 | $127 – $133 |
| Coding Task | 3,000 / 2,000 | $135 – $154 |
| Invoice Processing | 1,500 / 600 | $56 – $62 |
| Call Summary | 2,000 / 700 | $76 – $79 |
| Agent Workflow | 6,000 / 3,000 | $215 – $270 |
Price history
DeepSeek V3.2 has been in our daily price archive since tracking began on 2026-08-15, when it listed at $0.269 input / $0.40 output per 1M tokens. The price we quote has changed 7 times since then, 4 of them a change of quoted host rather than a repricing. The latest change, on 2026-11-01, was -9.0% on the blended price (30% input, 70% output), to $0.26 input / $0.38 output (quoted host changed, not a repricing). Net of all changes, the blended price is -4.6% against 2026-08-15.
| Date | Old price (in / out) | New price (in / out) | Blended change |
|---|---|---|---|
| 2026-11-01 | $0.28 / $0.42 | $0.26 / $0.38 | -9.0% (quoted host changed, not a repricing) |
| 2026-10-17 | $0.259 / $0.80 | $0.28 / $0.42 | -40.7% (quoted host changed, not a repricing) |
| 2026-10-10 | $0.259 / $0.42 | $0.259 / $0.80 | +71.6% |
| 2026-10-08 | $0.28 / $0.42 | $0.259 / $0.42 | -1.7% (quoted host changed, not a repricing) |
| 2026-09-28 | $0.269 / $0.40 | $0.28 / $0.42 | +4.8% (quoted host changed, not a repricing) |
2 earlier changes are not listed.
Alternatives to DeepSeek V3.2
None of the 4 comparable models below costs less than DeepSeek V3.2 on the blended price.
- Next cheaper model in the Budget price tier: Gemma 4 31B (Google) at $0.34 blended, 1% less.
- Next more expensive: Particle-1 (particle-ai) at $0.365 blended, 6% more.
The nearest models in the Budget price tier on Arena ELO, each with its blended price and score difference against DeepSeek V3.2.
- Beam 1.0 501B A23B (reflection): $0.30 in / $1.00 out per 1M, Arena ELO 1425 (130% more expensive, same Arena ELO)
- MiniMax M3 (MiniMax): $0.30 in / $1.20 out per 1M, Arena ELO 1428 (170% more expensive, 3 ELO points higher)
- DeepSeek V3.2 Exp (DeepSeek): $0.27 in / $0.41 out per 1M, Arena ELO 1422 (7% more expensive, 3 ELO points lower)
- Mistral Large 3 2512 (Mistral): $0.50 in / $1.50 out per 1M, Arena ELO 1429 (249% more expensive, 4 ELO points higher)
- Cheapest within 15 Arena points: Gemma 4 26B A4B (Google), 73% cheaper, 9 ELO points higher, out of 40 models in that band.
- Highest Arena ELO within 20% of this price: DeepSeek V4.1 Flash (DeepSeek), 17% cheaper, 38 ELO points higher, out of 14 models in that price band.
DeepSeek V3.2 head to head
Other models from DeepSeek
The 8 nearest to DeepSeek V3.2 in price, of 13 other DeepSeek models in the catalogue. Prices per 1M tokens, input then output.
- DeepSeek V4 Flash 0423: $0.09 / $0.18
- DeepSeek V4 Flash 0731: $0.0434 / $0.30
- DeepSeek V4.1 Flash: $0.022 / $0.40
- DeepSeek V3.2 Exp: $0.27 / $0.41
- DeepSeek V4.1 Flash 1017: $0.15 / $0.60
- DeepSeek V3: $0.32 / $0.89
- DeepSeek V3.1: $0.25 / $0.95
- DeepSeek V3.1 Terminus: $0.27 / $1.00
All 14 DeepSeek models and prices
Common questions about DeepSeek V3.2 pricing
- How much does DeepSeek V3.2 cost?
- DeepSeek V3.2 costs $0.26 per 1M input tokens and $0.38 per 1M output tokens at DeepInfra, the host whose rate we quote. DeepInfra serves a quantized fp4 build, which is not the same product as the full-precision model. Prices as listed on 2026-11-12.
- Which host is cheapest for DeepSeek V3.2?
- Cheapest host: GMICloud at $0.2088 input / $0.3096 output per 1M tokens, 19% below the price we quote on the blended price (30% input, 70% output). It serves a quantized fp8 build, which is not the same product as the full-precision model. A support ticket workload at 100K requests a month costs $46.80 there against $58.00 at the price we quote, a difference of $11.20 a month. The dearest, DigitalOcean, charges 2.7x the cheapest. Host rates as listed on 2026-11-12.
- Does DeepSeek V3.2 offer batch or cached pricing?
- Yes, for prompt caching only. Cache reads cost $0.13 per 1M tokens, as listed on 2026-11-12. No batch pricing is published.
- What would DeepSeek V3.2 cost per month?
- With the Support Ticket profile (1,500 input and 500 output tokens per request), 100,000 requests a month cost about $58 on DeepSeek V3.2 at the list price quoted on 2026-11-12.
- Can DeepSeek V3.2 be self-hosted?
- Yes. DeepSeek V3.2 is open source: its weights are published under MIT, an OSI-approved licence with no field-of-use limit, so it can be self-hosted as well as bought by the token.
- Has the price of DeepSeek V3.2 changed?
- DeepSeek V3.2 has been in our daily price archive since tracking began on 2026-08-15, when it listed at $0.269 input / $0.40 output per 1M tokens. The price we quote has changed 7 times since then, 4 of them a change of quoted host rather than a repricing. The latest change, on 2026-11-01, was -9.0% on the blended price (30% input, 70% output), to $0.26 input / $0.38 output (quoted host changed, not a repricing). Net of all changes, the blended price is -4.6% against 2026-08-15.
Its weights are published, which changes what you may run yourself: the cost gap between open weights and proprietary APIs.
These are list prices. Your real cost also depends on caching and batching, negotiated discounts, marketplace agreements, and the contracts you already have. OptimNow audits exactly that.
Build the business case · Get a FinOps review
LLM pricing · Models by maker · Cloud compute pricing · Compare models · Price barometer · FinOps guides · API and MCP server