DeepSeek V3.2 API pricing
As of 5 August 2026, DeepSeek V3.2 costs $0.28 per 1M input tokens and $0.40 per 1M output tokens, and scores 74.2% on coding benchmarks (best of Aider polyglot and SWE-bench Verified).
Prices updated 5 August 2026
DeepSeek V3.2 price by host
The same model weights are served by 2 providers. Prices are USD per 1M tokens; the cheapest output price is highlighted.
| Provider | Input / 1M | Output / 1M | Context |
|---|---|---|---|
| DeepSeek | $0.28 | $0.40 | 164K |
| Fireworks AI | $0.56 | $1.68 | 164K |
Coding: 74.2%
Coding via Aider polyglot & SWE-bench Verified; reasoning & math via Epoch AI.
- How much does the DeepSeek V3.2 API cost?
- As of 5 August 2026, DeepSeek V3.2 costs $0.28 per 1M input tokens and $0.40 per 1M output tokens, and scores 74.2% on coding benchmarks (best of Aider polyglot and SWE-bench Verified).
- What's the cheapest way to run DeepSeek V3.2?
- DeepSeek V3.2 is served by 2 hosts. The lowest output price is $0.40 per 1M tokens via DeepSeek — the same weights, so switching host is a like-for-like saving.
Know what you're really spending on DeepSeek V3.2.
StackSpend connects read-only to your AI providers, normalises every model into one daily signal, and flags cheaper equal-quality swaps from this same dataset.