DeepSeek-V3.2 (Thinking): Benchmarks, Pricing & Context Window
DeepSeek-V3.2 (Thinking) is a language model from DeepSeek, released in December 2025.
DeepSeek-V3.2 in thinking mode. A powerful reasoning model with 685B parameters using DeepSeek Sparse Attention (DSA). This mode enables extended chain-of-thought reasoning for complex problem-solving tasks. Supports JSON output and tool
DeepSeek-V3.2 (Thinking) benchmarks
Capability tiers
Standing within each category, adjusted for leaderboard depth.
Real tasks performance
High-confidence performance for DeepSeek-V3.2 (Thinking) across real-world prompt categories. Only 95% intervals at most 4 points wide are shown.
Performance by conversation depth
How DeepSeek-V3.2 (Thinking) holds up as conversations get longer.
Quality Tracker
DeepSeek-V3.2 (Thinking) Performance Across Datasets
Scores sourced from the model's scorecard, paper, or official blog posts
llm-stats.com - Wed Sep 09 2026DeepSeek-V3.2 (Thinking) pricing
Providers
DeepSeek-V3.2 (Thinking) starts at $0.280 per million input tokens and $0.420 per million output tokens via DeepSeek.
| Provider | Input $/M | Cached input $/M | Output $/M | Context in / out | TTFT p95 s | Output p5 c/s | Modalities in / out |
|---|---|---|---|---|---|---|---|
 DeepSeek | $0.280 | — | $0.420 | 131.1K/65.5K | 0.50 | — | / |
Cached input is the discounted price for prompt tokens served from a provider cache. TTFT is time to first token. Output is characters per second; p5 is the sustained floor exceeded by 95% of observed requests.
DeepSeek-V3.2 (Thinking) model size
DeepSeek-V3.2 (Thinking) has 685 billion parameters. See how it compares to other models in the same parameter range.
DeepSeek-V3.2 (Thinking) context window
Input and output token limits for DeepSeek-V3.2 (Thinking), plus how it ranks on long-context understanding.
Try now
huggleMake it with
DeepSeek-V3.2 (Thinking).
DeepSeek-V3.2 (Thinking) latency
DeepSeek-V3.2 (Thinking) time to first token, sustained output throughput, and failed-request rate from live model usage over the trailing 7 days.
DeepSeek-V3.2 (Thinking) examples
Recent arena outputs from DeepSeek-V3.2 (Thinking), picked from the highest-ranked matchups.
DeepSeek-V3.2 (Thinking) license
DeepSeek-V3.2 (Thinking) is released under the MIT license, which permits commercial use, has 685.0B parameters.
- License
- MIT
- Commercial use allowed
- Parameters
- 685.0B
MIT License - allows commercial use
DeepSeek-V3.2 (Thinking) resources
Official sources for DeepSeek-V3.2 (Thinking): provider documentation, official playground, paper or system card, model weights.
DeepSeek-V3.2 (Thinking) vs other models
The most-compared alternatives to DeepSeek-V3.2 (Thinking) are Grok-3, Seed 2.0 Lite, MAI-Thinking-1. Open any pair side-by-side for benchmarks, pricing, context, and latency.
Models like DeepSeek-V3.2 (Thinking)
Models ranked just above and below DeepSeek-V3.2 (Thinking) by LLM Stats score.
FAQ
Common questions about DeepSeek-V3.2 (Thinking).