DeepSeek · Open weights
DeepSeek V3.2
DeepSeek V3.2 (non-thinking mode) — official successor to V3.2-Exp. Reasoning-first model built for agents with GPT-5 level performance. Balanced inference vs. output length for everyday use. First DeepSeek model with thinking-in-tool-use capability. FP8.
- Tools
- Structured output
- PDF input
- Input
- Text
- Output
- Text
- Discovery tier
- Open weights
Pricing
USD per 1M tokens- Input
- $0.21
- Output
- $0.31
- Cache read
- $0.0216
Reference prices as published by the model provider. API example
Specifications
- Context window
- 163K
- Max output
- 65.5K
- Average output (7 days)
- 408 tokens
- Longer than 41% of 473 models
Side by side
Compare models
Put DeepSeek V3.2 next to up to three other models. The comparison stays in the link, so you can share it.
Public evaluations
Benchmarks
Independent evaluation results, with how each score ranks among the models in the Noumen catalog.
Intelligence Index
16.0
Better than 40% of 326 scored models
Indexes
Math Index
59.0Composite of the math evaluations
Better than 56% of 118 scored models
Document reasoning
AA-LCRv1.1
45.7%Reasoning across long documents
Better than 23% of 309 scored models
Reasoning and knowledge
Humanity’s Last Exam
11.2%Expert-level questions across academic fields
Better than 36% of 323 scored models
GPQA Diamond
75.1%Graduate-level science questions
Better than 36% of 291 scored models
MMLU-Pro
83.7%Multi-subject knowledge and reasoning
Better than 78% of 132 scored models
IFBench
49.0%Following precise output instructions
Better than 48% of 215 scored models
Coding
LiveCodeBench
59.3%Recent competitive programming problems
Better than 55% of 129 scored models
Math
AIME 2025
59.0%Competition mathematics, 2025 problems
Better than 56% of 118 scored models
Legacy evaluations
Terminal-Bench Hard
32.6%Earlier terminal task suite
Better than 60% of 212 scored models
τ²-Bench Telecom
78.9%Tool use in a customer-support simulation
Better than 58% of 214 scored models
AA-LCR (unversioned)
45.7%Long-context reasoning, earlier grading
Better than 23% of 309 scored models
Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis
Live routing data
Performance and reliability
Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.
- Throughput
- 48.4 tok/s
- Faster than 35% of 256 models
- Time to first token
- 2.5 s
- Quicker than 63% of 256 models
- Prompt cache hits
- 48.6%
- About $0.12 per 1M input tokens with caching
- Request success (24h)
- 99.7%
- Reliable
Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.
- Healthy (98% or more)
- Degraded (90–98%)
- Down (under 90%)
- No requests
Developers
Use it through the API
Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.
curl https://www.noumen.xyz/api/v1/chat/completions \
-H "Authorization: Bearer $NOUMEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek/deepseek-v3.2",
"messages": [{ "role": "user", "content": "Explain this stack trace." }]
}'import os
from openai import OpenAI
client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
model="deepseek/deepseek-v3.2",
messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)DeepSeek
Related models
Other DeepSeek models, starting with the same family.