Atlas / Models / DeepSeek / DeepSeek: DeepSeek V3.2 Speciale

DeepSeek: DeepSeek V3.2 Speciale✓ Catalog verified

deepseek/deepseek-v3.2-speciale

DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance. It builds on DeepSeek Sparse Attention (DSA) for efficient long-context processing, then scales post-training reinforcement learning to push capability beyond the base model. Reported evaluations place Speciale ahead of GPT-5 on difficult reasoning workloads, with proficiency comparable to Gemini-3.0-Pro, while retaining strong coding and tool-use reliability. Like V3.2, it benefits from a large-scale agentic task synthesis pipeline that improves compliance and generalization in interactive environments.

Input price
$0.40 /1M
Output price
$1.20 /1M
Context
164K
Modalities
text
Released
Dec 1, 2025
Tool calling
Atlas signal
85/100
01

Overview

DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance. It builds on DeepSeek Sparse Attention (DSA) for efficient long-context processing, then scales post-training reinforcement learning to push capability beyond the base model. Reported evaluations place Speciale ahead of GPT-5 on difficult reasoning workloads, with proficiency comparable to Gemini-3.0-Pro, while retaining strong coding and tool-use reliability. Like V3.2, it benefits from a large-scale agentic task synthesis pipeline that improves compliance and generalization in interactive environments.

Access: available through the official DeepSeek API. Context window: 164K.

Pricing and metadata from the ModelsAtlas catalog.Last refreshed Aug 5, 2026
02

Specifications

API identifierdeepseek/deepseek-v3.2-speciale
ProviderDeepSeek
Model typeText LLM
Context window164K catalog
Input modalitiestext
Output modalitiestext
ReleasedDec 1, 2025
TokenizerDeepSeek
ModeratedNo
Architecture modalitytext->text
Supported parameters
frequency_penaltyinclude_reasoninglogit_biasmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetop_ktop_p
Source: provider documentation + catalog feedMethodology →

API defaults

Default parameters
{
  "top_p": 0.95,
  "temperature": 1,
  "frequency_penalty": null
}
03

Pricing

Live pricing components for deepseek/deepseek-v3.2-speciale as published in the catalog.

$0.40
Input /1M
$1.20
Output /1M
Per request
Per image
Pricing componentRaw unit priceNormalizedUnit
Prompt tokens$4e-7$0.40 / 1MPer input token
Completion tokens$0.0000012$1.20 / 1MPer output token
Request feeN/AN/APer request
Image feeN/AN/APer image unit
Web search feeN/AN/APer search request
Source: catalog pricing feedCompare all pricing →Cheapest models →
04

Cost calculator

Estimate monthly spend from your own token volumes.

Scale / Volume

Estimated Total Cost

$0.44Calculated from current list pricing in the ModelsAtlas catalog.
05

Capabilities

CapabilityStatusWhat it means
Visual UnderstandingNot advertisedImage and document analysis support
Audio ProcessingNot advertisedSpeech and voice aligned flows
Tool CallingNot advertisedSupports tools / function calling
Self-Hosting✓ SupportedDeploy outside managed APIs
Derived from catalog capability tags and supported parameters
06

Capability signals

Directional signals derived from model metadata and capability tags — not official benchmark submissions.

DeepSeek: DeepSeek V3.2 Speciale
MMLU Signal (Reasoning)
97signal
Coding Signal (HumanEval proxy)
82signal
Math Signal (GSM8K proxy)
83signal
Science Signal (GPQA proxy)
79signal
Estimated from metadata — not an official benchmark runAll benchmarks →Methodology →
07

Quick start

Call DeepSeek: DeepSeek V3.2 Speciale through an OpenAI-compatible client.

from openai import OpenAI

client = OpenAI(
    base_url="https://openrouter.ai/api/v1",
    api_key="YOUR_API_KEY"
)

response = client.chat.completions.create(
    model="deepseek/deepseek-v3.2-speciale",
    messages=[{"role": "user", "content": "Explain quantum physics."}]
)

print(response.choices[0].message.content)
08

Alternatives

Closest models by context window from a different provider.

09

Sources & attribution

Pricing and metadata are maintained in the ModelsAtlas catalog.

Capability bars use metadata tags — directional estimates onlyHow we source and verify data →
10

Frequently asked questions

How much does DeepSeek: DeepSeek V3.2 Speciale cost?

$0.40 per 1M input tokens and $1.20 per 1M output tokens on the official DeepSeek API.

What is the context window of DeepSeek: DeepSeek V3.2 Speciale?

DeepSeek: DeepSeek V3.2 Speciale supports up to 164K tokens of context.

Does DeepSeek: DeepSeek V3.2 Speciale support tool / function calling?

DeepSeek: DeepSeek V3.2 Speciale does not advertise first-class tool/function calling. Check vendor docs for the latest capabilities.

How do I access DeepSeek: DeepSeek V3.2 Speciale?

Use the official DeepSeek API with the model id `deepseek/deepseek-v3.2-speciale`. See the Quick start section above for code examples.