DeepSeek API Pricing (2026)
DeepSeek is a Chinese AI lab offering high-performance open-source models at extremely competitive prices, with off-peak discounts of 50-75%. Founded in 2023 and headquartered in Hangzhou, China, DeepSeek offers 2 active models through their API. All pricing data below is sourced directly from the DeepSeek pricing page and is updated regularly.
Prices verified Mar 6, 2026
DeepSeek Model Pricing — All Models
All prices are in USD per 1 million tokens ($/M tokens). Input tokens are the text you send to the model; output tokens are the generated response.
| Model | Tier | Input | Output | Cached Input | Batch Input | Batch Output |
|---|---|---|---|---|---|---|
DeepSeek V3.2 DeepSeek's flagship model updated to V3.2 with unified pricing and excellent performance at extremely competitive rates. | Premium | $0.28/M | $0.42/M | $0.03/M | — | — |
DeepSeek R1 DeepSeek's reasoning model with V3.2 architecture, offering frontier reasoning at a fraction of competitors' costs. | Reasoning | $0.28/M | $0.42/M | $0.03/M | — | — |
Prices last verified: 2026-03-06. Always confirm on the official pricing page before production use.
DeepSeek Model Capabilities
Key technical capabilities across all DeepSeek models. Use this table to identify which models support the features your application requires.
| Model | Context | Max Output | Vision | Functions | JSON Mode | Caching | Batch API | Reasoning |
|---|---|---|---|---|---|---|---|---|
| DeepSeek V3.2 | 128,000 | 8,192 | — | ✓ | ✓ | ✓ | — | — |
| DeepSeek R1 | 128,000 | 64,000 | — | — | ✓ | ✓ | — | ✓ |
When to Choose DeepSeek
Cost-sensitive workloads
Use DeepSeek V3.2 — DeepSeek's most affordable model at $0.28/M input / $0.42/M output. Estimated cost at 100K requests/month: $49.00.
Complex or high-accuracy tasks
Use DeepSeek R1 — DeepSeek's most capable model (reasoning tier) with a 128,000-token context window. Priced at $0.28/M input / $0.42/M output.
High-volume pipelines
DeepSeek models are well-suited for high-throughput workloads. Consider batch API (where supported) for an additional 50% cost reduction on asynchronous pipelines. Check the pricing table above for batch pricing availability.
Compare DeepSeek Models vs Competitors
See detailed cost comparisons between DeepSeek models and models from other providers, including side-by-side pricing tables and monthly cost projections.
Use Cases for DeepSeek Models
Browse use case guides where DeepSeek models are recommended, including cost-effective model rankings and monthly cost estimates for each workload.
Frequently Asked Questions: DeepSeek API Pricing
What is the cheapest DeepSeek model?
The cheapest DeepSeek model by input token price is DeepSeek V3.2, priced at $0.28/M for input tokens and $0.42/M for output tokens. It is best suited for high-volume, cost-sensitive workloads where speed and price matter most.
What is the most capable DeepSeek model?
DeepSeek R1 is DeepSeek's most capable model, classified as a reasoning-tier model. It supports a context window of 128,000 tokens and is priced at $0.28/M input / $0.42/M output per million tokens.
How does DeepSeek API pricing work?
DeepSeek charges per token consumed — separately for input tokens (the text you send) and output tokens (the text the model generates). Prices are listed in USD per 1 million tokens ($/M). DeepSeek was founded in 2023 and is headquartered in Hangzhou, China. All pricing shown here is sourced from the official DeepSeek pricing page.
Does DeepSeek offer batch API pricing?
DeepSeek's current model lineup does not include batch API pricing at this time. For cost reduction on high-volume workloads, consider prompt caching (where supported) or running multiple concurrent requests with standard pricing.
Which DeepSeek models support prompt caching?
2 DeepSeek models support prompt caching: DeepSeek V3.2, DeepSeek R1. Prompt caching stores repeated context (system prompts, knowledge base content) in memory and applies a discount to cached input tokens, making it highly effective for applications that reuse the same context across many requests.
Do DeepSeek models support vision / image inputs?
None of the current DeepSeek models in our dataset support vision input. DeepSeek models are text-focused. For multimodal workloads requiring image understanding, consider comparing models from other providers using our comparison tool.
DeepSeek pricing data last verified: 2026-03-06. Prices may change — verify on the official DeepSeek pricing page before production use.