DeepSeek-V3 (Mar 2025)
DeepSeekOpen weights
DeepSeek-V3 (Mar 2025) by DeepSeek costs — per 1M input tokens and — per 1M output tokens, with a —-token context window.
Pricing
Input (per 1M tokens)
—
Output (per 1M tokens)
—
Cached input (per 1M)
—
Specifications
- Provider
- DeepSeek
- Context window
- —
- Parameters
- 671B
- Released
- Mar 2025
- Open weights
- Yes
- Frontier model
- No
Benchmarks
- Aider Polyglot
- 55.1
Compare DeepSeek-V3 (Mar 2025) with…
DeepSeek-V3 (Mar 2025) vs DeepSeek V3.1 Terminus$1.00/1MDeepSeek-V3 (Mar 2025) vs Qwen2.5 Coder 32B Instruct$1.00/1MDeepSeek-V3 (Mar 2025) vs Nemotron 3 Super 120B A12B$1.00/1MDeepSeek-V3 (Mar 2025) vs MiniMax M2$1.02/1MDeepSeek-V3 (Mar 2025) vs MiniMax M2.5$0.95/1MDeepSeek-V3 (Mar 2025) vs DeepSeek V3.1$0.95/1M
FAQ
How much does DeepSeek-V3 (Mar 2025) cost?
DeepSeek-V3 (Mar 2025) is priced at — per 1M input tokens and — per 1M output tokens. Output tokens cost more than input.
What is DeepSeek-V3 (Mar 2025)'s context window?
DeepSeek-V3 (Mar 2025) supports a context window of — tokens — the maximum prompt plus response length it can handle in a single request.
Is DeepSeek-V3 (Mar 2025) open weight?
Yes — DeepSeek-V3 (Mar 2025) has openly available weights.
Data as of 2026-10-16. Source: Epoch AI, Data on AI Models (CC BY 4.0), epoch.ai. Pricing is per 1M tokens (USD); confirm with the provider before production use.