DeepSeek V3.2 / チャット
The official version of DeepSeek-V3.2 balances reasoning capability and output length, making it suitable for daily use—such as question-answering scenarios and general Agent task scenarios.
Playground Chat
AIモデルとの会話を開始します。何でも質問できます。
AI生成結果の正確性は異なる場合があります。
DeepSeek V3.2Deep Reasoning with OpenAI Protocol
DeepSeek V3.2 delivers configurable deep thinking, adjustable reasoning depth, token probability signals, and parallel tool calling through a drop-in OpenAI-compatible chat API.
At a Glance
Reasoning & Tool Architecture
Enterprise-ready chat with deep thinking controls.
Minimal → High Effort Ladder
Dial reasoning depth per request — Minimal for speed, High for multi-step proofs and analysis.
Confidence via Logprobs
Inspect token probabilities and run parallel tools for agent evaluation pipelines.
Streaming for Long Analyses
Stream reasoning and final tokens so complex analyses feel progressive instead of blocking on a single long wait.
Sampling for Tone Control
Temperature and top-p (plus penalties) let teams match brand voice without rewriting system prompts every time.
How It Works
A production-ready path from brief to finished asset.
Enable Thinking
Keep thinking on for analytical tasks; disable for pure extraction or rewrite jobs.
Set Effort
Minimal for speed, High for multi-document synthesis or formal proofs.
Stream Reasoning Tokens
Surface intermediate tokens when you want visible progress in the UI.
Evaluate with Logprobs
Collect probability signals for offline scoring and regression tests.
Deep Reasoning Domains
Where analytical depth wins.
Technical Analysis
High effort for multi-hop technical reasoning.
Agent Tool Loops
Parallel tools with logprob-based confidence.
Document QA
Structured answers from long documents.
Code Review
Medium/high effort for defect finding.
Prompt & Usage Tips
Practical guidance for reliable first-pass results.
When you need reasoning, request intermediate steps explicitly so effort is spent usefully.
Put documents/context first, then the instruction. Models follow that order more reliably.
High depth is for hard problems. Everyday Q&A is faster at Minimal or Low.
DeepSeek Quickstart
Stream reasoning responses with tool support.
curl -X POST "https://api.powertokens.ai/v1/chat/completions" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v3-2-251201",
"messages": [
{"role": "user", "content": "Walk through the proof of the intermediate value theorem."}
],
"stream": true,
"temperature": 0.7
}'Technical Specifications
Confirmed parameters and runtime execution protocols.
料金詳細
このモデルの実際の課金は、API リクエストで渡される特定のパラメータに基づいて動的に計算されます。以下は具体的な組み合わせとそれに対応する料金です：
| モダリティ | 入力クレジット | 出力クレジット | 入力価格 | 出力価格 | 暗黙的キャッシュヒット | 明示的キャッシュヒット | キャッシュ作成 |
|---|---|---|---|---|---|---|---|
| 0 - 32K | 280/ 1M Tokens | 420/ 1M Tokens | $0.280 | $0.420 | $0.056 56/ 1M Tokens | -- | -- |
| 32K - 128K | 560/ 1M Tokens | 840/ 1M Tokens | $0.560 | $0.840 | $0.056 56/ 1M Tokens | -- | -- |
Models from the Same Channel
Explore complementary models and alternative versions from the same provider channel.
DeepSeek-V4.1-Flash
deepseek-v4-1-flash-260910
Lightweight flagship of DeepSeek's new Causal Encoder-Decoder family. 552B-parameter MoE with asymmetric activation—8B for input, 16B for output—delivering flagship-level intelligence at significantly lower inference cost . Native multimodal vision understanding. 1M-token context window with 384K max output. KV cache compressed to 1/4 HBM and 1/8 SSD of previous generation . Outperforms V4-Pro on Terminal-Bench 2.1 (90.6), DeepSWE v1.1 (74.2), and CyberGym (88.1)
Seed 2.0 Pro
seed-2-0-pro-260328
Focused on long-chain reasoning and stability in complex task execution, designed for complex real-world business scenarios.
Seed 2.0 Lite
seed-2-0-lite-260228
Balances generation quality and response speed, making it a strong general-purpose production model.
Seed 2.0 Mini
seed-2-0-mini-260215
Built for low-latency, high-concurrency, cost-sensitive use cases, with flexible deployment, four-tier thinking, and multimodal understanding.
Recommended Related Models
Explore complementary video and multimodal models with your unified API key.
GLM-5.3
glm-5.3
Zhipu AI's flagship text model optimized for complex software engineering and long-horizon Agent tasks. Features a 1M-token context window with mandatory reasoning (3 levels: low/high/max). Coding capability improved 50% over GLM-5.2 on Z.ai Code Bench; scores SOTA on Terminal Bench 3.0. Excels in cybersecurity tasks (cyber vulnerability discovery)
MiniMax-M3.1-Flash
MiniMax-M3.1-Flash
Next-gen efficiency-focused multimodal model (428B MoE) for agentic workflows, coding, and computer use . Features a 1M-token context window and tunable thinking depth (low–max) . Scores 73.7% on DeepSWE v1.1, 61.7 on OSWorld V2, and 82% on CyberGym, while delivering 200+ TPS output speed at 70% lower agent-task cost than its predecessor
GLM-5.3 Flash
glm-5.3-flash
Z.ai's native multimodal model (320B-A18B) with text, image, video, and file input. Features hybrid sparse + linear attention, reducing attention compute 3x and KV cache 4.4x vs. GLM-5.3. Scores 57 on Artificial Analysis Intelligence Index—matching Claude Opus 4.8—with 1M context and 128K max output. Excels in agentic coding, vision-based Computer Use, and CAD reproduction
Qwen3.8 Flash
qwen3.8-flash
Alibaba's cost-efficient multimodal reasoning model. Supports text, image, and video inputs with text output. Features a native 1M-token context window for long documents, codebases, and agentic workflows. Excels in coding assistance, desktop interaction, chart analysis, and long-video understanding. Compatible with OpenAI/Anthropic protocols for seamless integration
Frequently Asked Questions
Everything you need to know before integrating this model.
Start Building with DeepSeek V3.2 Today
Create an account in seconds to receive free test credits. No upfront payment or long-term contract required.