Model

DeepSeek V3.2

Name: DeepSeek V3.2
Brand: deepseek

Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use

35 providers · 40 listingsreleased Dec 1, 2025knowledge cutoff 2024-07open weights

$0.18

lowest input /M · CrofAI

$0.326

lowest output /M · TokenGo

164K

context window

164K

max output

Compare providers (sorted by listed input price)

32 of 35 providers publish a price for DeepSeek V3.2. The cheapest full-context option is CrofAI at $0.18/$0.35 per M. Listed prices span 4.6× for identical weights.

			Cache read /M	Cache write /M		Capabilities
CrofAI deepseek-v3.2	$0.18	$0.35	$0.04	—	164K	no reasoningtoolsno structuredno vision	Jul 22, 2025
TokenGo deepseek/deepseek-v3.2	$0.217	$0.326	$0.06	—	128K	reasoningtoolsstructuredno vision	Dec 1, 2025
LLM Gatewayvariant deepinfra/deepseek-v3.2	$0.26	$0.38	$0.13	—	160K	no reasoningtoolsno structuredno vision	Dec 1, 2025
DevPass (LLM Gateway) deepseek-v3.2	$0.26	$0.38	$0.13	—	164K	reasoningtoolsstructuredvision	Sep 29, 2025
Deep Infra deepseek-ai/DeepSeek-V3.2	$0.26	$0.38	$0.13	—	164K	reasoningtoolsstructuredno vision	Dec 2, 2025
Meganova deepseek-ai/DeepSeek-V3.2	$0.26	$0.38	—	—	164K	no reasoningtoolsstructuredno vision	Dec 3, 2025
LLM Gatewayvariant novita/deepseek-v3.2	$0.269	$0.4	$0.135	—	164K	reasoningtoolsno structuredno vision	Dec 1, 2025
NovitaAI deepseek/deepseek-v3.2	$0.269	$0.4	$0.135	—	164K	reasoningtoolsstructuredno vision	Dec 1, 2025
Kilo Gateway deepseek/deepseek-v3.2	$0.269	$0.4	$0.135	—	164K	reasoningtoolsstructuredno vision	Dec 1, 2025
OpenRouter deepseek/deepseek-v3.2	$0.269	$0.4	$0.135	—	164K	reasoningtoolsstructuredno vision	Dec 1, 2025
SiliconFlow (China)variant deepseek-ai/DeepSeek-V3.2	$0.27	$0.42	—	—	164K	reasoningtoolsstructuredno vision	Dec 3, 2025
SiliconFlow (China)variant Pro/deepseek-ai/DeepSeek-V3.2	$0.27	$0.42	—	—	164K	reasoningtoolsstructuredno vision	Dec 3, 2025
Alibaba (China) siliconflow/deepseek-v3.2	$0.27	$0.42	—	—	164K	reasoningtoolsstructuredno vision	Dec 3, 2025
Poe novita/deepseek-v3.2	$0.27	$0.4	$0.13	—	128K	reasoningtoolsno structuredvision	Dec 1, 2025
Helicone deepseek-v3.2	$0.27	$0.41	—	—	164K	no reasoningtoolsno structuredno vision	Sep 22, 2025
Abacus deepseek-ai/DeepSeek-V3.2	$0.27	$0.4	—	—	128K	reasoningtoolsno structuredno vision	Jun 15, 2025
SiliconFlow deepseek-ai/DeepSeek-V3.2	$0.27	$0.42	—	—	164K	reasoningtoolsstructuredno vision	Dec 3, 2025
NanoGPTvariant deepseek/deepseek-v3.2	$0.28	$0.42	$0.14	—	163K	no reasoningtoolsstructuredvision	Dec 1, 2025
Vivgrid deepseek-v3.2	$0.28	$0.42	—	—	128K	reasoningtoolsno structuredno vision	Dec 1, 2025
LLM Gatewayvariant bytedance/deepseek-v3.2	$0.28	$0.42	$0.056	—	131K	reasoningtoolsno structuredno vision	Dec 1, 2025

Best price history

Daily lowest published price across live providers. A lower line is better for the buyer.

Best input /M: $0.18unchanged
Best output /M: $0.35unchanged
Provider coverage: 40; 40 at the start · 24 snapshots

Published best prices are unchanged.

24 daily snapshots collected from Aug 21, 2026 through Sep 14, 2026. A price chart will appear after the first real move.

Price badge

![DeepSeek V3.2 price](https://www.vaanalytics.in/badge/deepseek/deepseek-v3.2.svg)

https://www.vaanalytics.in/badge/deepseek/deepseek-v3.2.svg

Live SVG, regenerated on every hourly sync — paste it into a README to always show the current cheapest listed price. Updates when prices change; no build step needed on your side.

Recent changes

Changelog →

capabilityvia amazon-bedrock · Sep 13, 2026
- name: DeepSeek-V3.2 → DeepSeek V3.2
- description: DeepSeek chat model for instruction following, coding, and analysis → Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use
- release_date: 2026-02-06 → 2025-12-01
repricedvia vercel · Sep 11, 2026
- input: $0.28 → $0.62
- output: $0.42 → $1.85
- cache_read: $0.028 → —
repricedvia openrouter · Aug 30, 2026
- output: 66K → 147K tokens
- input: $0.269 → $0.26
- output: $0.4 → $0.38
- cache_read: $0.135 → $0.13
contextvia kilo · Aug 30, 2026
- output: 66K → 147K tokens
contextvia kilo · Aug 29, 2026
- output: 66K → 147K tokens
repricedvia openrouter · Aug 29, 2026
- output: 66K → 147K tokens
- input: $0.269 → $0.26
- output: $0.4 → $0.38
- cache_read: $0.135 → $0.13

Weights

Permissivemit685B params

Standard open-source terms. Commercial use without additional conditions. Weights download without an access request.

Derived from deepseek-ai/DeepSeek-V3.2-Exp-Base.

licence and gating from the Hugging Face model card ↗ — the card is authoritative, this is a summary

Hugging Face

Prices are per million tokens (USD). “—” means the provider does not publicly list a price for this model. Data from models.dev; verify with the provider before purchasing.