DeepSeek V3.2 Speciale

Provider: DeepSeek CN

No longer listed on OpenRouter since 2026-05-29. The figures below are the last recorded values.
Input price
$0.29
/ 1M tokens
Output price
$0.43
/ 1M tokens
Cache price
$0.058
/ 1M tokens (read)
Context
164K tokens
Max output 164K tokens

Overview

DeepSeek V3.2 Speciale is a specialized mid-tier model from DeepSeek (CN) designed for workloads requiring advanced reasoning and structured outputs. It features a substantial context window for processing and generating long, complex responses. This model targets applications needing high quality and specific formatting.

Strengths

  • Very large 163,840-token context window for extensive input and output.
  • Robust 'thinking' capability supports complex problem-solving and reasoning tasks.
  • Dedicated 'json-mode' for reliable structured data generation.
  • Highly competitive output pricing at $0.431 per million tokens compared to peers.
  • Offers an identical 163,840-token maximum output length.

Best use cases

  • Analyzing and summarizing lengthy legal documents or research papers with structured outputs.
  • Complex data extraction and transformation requiring advanced reasoning capabilities.
  • Developing sophisticated chatbots or conversational agents handling deep context.
  • Generating structured code or configuration files with specific JSON formatting.
  • Content creation that demands extensive context and consistent, formatted responses.

Pricing position

DeepSeek V3.2 Speciale's input price of $0.287/M tokens positions it at the higher end among DeepSeek's V3 models, but competitively with peers like MiniMax M2.1 ($0.29/M) and Gemini 2.5 Flash ($0.3/M). Its output price of $0.431/M tokens is significantly lower than many similarly-priced peer models (e.g., MiniMax M2.1 at $0.95/M, Gemini 2.5 Flash at $2.5/M), making it highly cost-effective for applications with high output token usage.

Caveats

This model does not offer a free tier, and its capabilities are limited to text processing, lacking features like vision or audio understanding. For tasks not requiring its specific 'thinking' or large context window, DeepSeek's V4 Flash ($0.112/M input) or V3.2 ($0.252/M input) could be more cost-efficient alternatives.

Notes drafted by an LLM from the figures on this page (2026-05-21) and rewritten when the price changes.

Key facts

Max output
164K tokens
Tracked since
2026-05-08
Last updated
2026-05-29

Features

JSON outputReasoning

Price history

  • 2026-05-29 No longer listed

Monthly cost estimate

Scale Input (M) Output (M) Monthly
0.1M (hobby) 0.1 0.05 $0.05
1M (small app) 1 0.5 $0.50
10M (mid scale) 10 5 $5.03
100M (large) 100 50 $50.25

※ Cache and batch discounts not factored in. Detailed calculator

Official resources

Other models from DeepSeek

Similar-priced models

Source: OpenRouter API