All models
DeepSeek
DeepSeek V3.2 fastruns locally
deepseek/deepseek-v3-2
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...
Input / 1M
$0.28
USD per million tokens
Output / 1M
$0.40
USD per million tokens
Est. cost / token
$0.00000031
$0.31 per 1M, 3:1 mix
Context window
164K
max output 164K
Major features
Open weightsCachingReasoningJSON schemaTools
Task suitability (heuristic 1–5, not a benchmark)
- coding3
- reasoning3
- math3
- summarization4
- extraction4
- classification4
- chat3
- creative3
- translation4
- vision1
- long context3
- agentic3
Run it locally
Will it run on my machine? →| Quantization | Quality | Memory needed (8K ctx) |
|---|---|---|
| Q2_K | poor | 309 GB |
| Q3_K_M | fair | 359 GB |
| Q4_K_Mrecommended | good | 447 GB |
| Q5_K_M | very good | 519 GB |
| Q6_K | excellent | 598 GB |
| Q8_0 | near-lossless | 772 GB |
| F16 | lossless | 1447 GB |
Estimates.
Price history
| Recorded | Input | Output | Cached | Source |
|---|---|---|---|---|
| 2026-11-05 03:20 | $0.28 | $0.40 | $0.028 | litellm+openrouter |
A row is added whenever a sync detects a price change. Vendor pricing source
Pricing (USD per 1M tokens)
- Input
- $0.28
- Output
- $0.40
- Cached input
- $0.028
- Blended (3:1)
- $0.31
Details
- Company
- DeepSeek
- Version
- 3.2
- Family
- DeepSeek
- Released
- 2025-12-01
- Parameters
- 685.4B
- Max output
- 164K
- Open weights
- Yes
- Inputs
- text
- Last synced
- 26 min ago