DeepSeek V3.2
by DeepSeektext->text12 endpoints
Neura Intelligence Index
22.6/ 100
Rank
#300—
Confidence
lowBenchmarks
3(8% coverage)
Overview
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...
Capabilities
Text generation
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
Top-P sampling
Stop sequences
Deterministic seed
Log probabilities
Modalities
Input
Text
Output
Text
Technical Specifications
Context Window
163.8K tokens
Max Output
65.5K tokens
Tokenizer
DeepSeek
Uptime (24h)
10000.0%
Parameters
671000000000B
Blended Cost
$0.4775/1M tokens
License
mit_+_model_license_(commercial_use_allowed)Open Weights
Country
🇨🇳CN
Model Family
DeepSeek V3
Release Date
December 25, 20241y
Supported Parameters
frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
Benchmark Performance
general
| Benchmark | Score | Source |
|---|---|---|
| SimpleQA | 24.9 | verified |
coding
| Benchmark | Score | Source |
|---|---|---|
| SWE-bench Verified | 42 | verified |
reasoning
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 59.1 | verified |
Performance Over Time
Pricing
Per 1M tokens
Input$0.28
Output$0.42
Cache read$0.03
Added
December 1, 2025
Last synced 10/23/2026