AgenticMarket
DeepSeekOpen WeightReleased: 2025-12-01V3.2
DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...
ARCHITECTURE & LIMITS
DeepSeekBENCHMARK EVALUATIONS & CODE METRICS
Verified scores from SWE-bench and Artificial Analysis.
Independent coding benchmark evaluating syntax, algorithms, and code completion
Broad multi-domain reasoning, science, and knowledge retrieval
Autonomous multi-step tool use, API calling, and task execution
TOKEN COST CALCULATOR
Estimate production API costs for V3.2.
COMMUNITY ENGINEERING EVALUATIONS
Verified ratings across 4 core software engineering dimensions.
Refactoring accuracy, syntax correctness, and adherence to design patterns.
Zero hallucinations, strict constraint compliance, and negative prompt respect.
Time-to-first-token (TTFT) and throughput tokens per second.
Token pricing economics relative to intelligence quality and output volume.
No text reviews submitted yet. Be the first developer to review V3.2!
RECOMMENDED AGENT SKILLS FOR V3.2
boardroom
/cs:boardroom <brief> — 6-phase multi-role deliberation across the C-suite with Phase 2 isolation, critic pre-screen, and synthesis. Outputs a board memo. Use when a decision spans multiple executive domains — e.g. a pricing change touching finance, positioning, and product, or a raise-vs-cut runway call.
FREQUENTLY ASKED QUESTIONS ABOUT V3.2
What are the input and output token prices for V3.2?+
V3.2 costs $0.27 per 1M input tokens and $0.40 per 1M output tokens. Cached input reads cost $0.13 per 1M tokens.
Does V3.2 support tool calling and agent execution?+
Yes, V3.2 natively supports tool calling, parallel function execution, and structured agent workflows.
What is the context window of V3.2?+
V3.2 features an active context window of 163,840 tokens with a maximum single-turn output of 65,536 tokens.
What AI Agent Skills work best with V3.2?+
Because V3.2 is optimized for agentic tool calling and code reasoning, it pairs well with Agent Skills like Boardroom, SEO Audit, and UI/UX Pro Max.