NOUMENOpen app
NOUMEN

DeepSeek · Open weights

DeepSeek V3.2 Exp

Added Jan 1, 2024Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

DeepSeek V3.2 Exp is an experimental V3-series model with improved long-context efficiency compared with earlier V3 releases. FP8.

Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.28
Output
$0.42

Reference prices as published by the model provider. API example

Specifications

Context window
163.8K
Max output
65.5K
Average output (7 days)
716 tokens
Longer than 62% of 468 models

Side by side

DeepSeek V3.2 Exp compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

DeepSeek V3.2 ExpDeepSeekThis modelDeepSeek V4.1 Flash 1017DeepSeekRemove
ProviderDeepSeekDeepSeek
Discovery tierOpen weightsOpen weights
AddedJan 1, 2024Oct 17, 2026
Limits
Context window163.84K1.05MBest
Max output65.54K393.22KBest
Average output716 tokens465 tokens
Pricing, USD per 1M tokens
Input$0.28$0.10Best
Output$0.42Best$0.60
Cache read—$0.003
Performance
Throughput17.9 tok/s263.1 tok/sBest
Time to first token3.7 s1.9 sBest
Request success (24h)99.8%99.8%
Benchmark indexes
Intelligence Index13.924.7Best
Math Index57.7—
Capabilities
ReasoningNoYes
ToolsNoYes
VisionNoYes
Structured outputNoYes
Reasoning effort—none, low, high, max

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • DeepSeek V3.2 Exp
  • DeepSeek V4.1 Flash 1017

Intelligence Index

  • DeepSeek V3.2 Exp13.9
  • DeepSeek V4.1 Flash 101724.7

DeepSeek V3.2 Exp: Better than 34% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Math Index

    Composite of the math evaluations

    • DeepSeek V3.2 Exp: 57.7
    • DeepSeek V4.1 Flash 1017: —

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • DeepSeek V3.2 Exp: —
    • DeepSeek V4.1 Flash 1017: 47.8%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • DeepSeek V3.2 Exp: —
    • DeepSeek V4.1 Flash 1017: 14.2%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • DeepSeek V3.2 Exp: —
    • DeepSeek V4.1 Flash 1017: 1,118 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • DeepSeek V3.2 Exp: —
    • DeepSeek V4.1 Flash 1017: 1,340 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • DeepSeek V3.2 Exp: —
    • DeepSeek V4.1 Flash 1017: 3.8%

  • AA-LCRv1.1

    Reasoning across long documents

    • DeepSeek V3.2 Exp: 44.7%
    • DeepSeek V4.1 Flash 1017: 55.3%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • DeepSeek V3.2 Exp: 9.0%
    • DeepSeek V4.1 Flash 1017: 10.8%Top

  • GPQA Diamond

    Graduate-level science questions

    • DeepSeek V3.2 Exp: 73.8%
    • DeepSeek V4.1 Flash 1017: —

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • DeepSeek V3.2 Exp: 83.6%
    • DeepSeek V4.1 Flash 1017: —

  • IFBench

    Following precise output instructions

    • DeepSeek V3.2 Exp: 43.1%
    • DeepSeek V4.1 Flash 1017: —

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • DeepSeek V3.2 Exp: —
    • DeepSeek V4.1 Flash 1017: 5.6%

  • SciCode

    Scientific computing in Python

    • DeepSeek V3.2 Exp: —
    • DeepSeek V4.1 Flash 1017: 35.5%

  • LiveCodeBench

    Recent competitive programming problems

    • DeepSeek V3.2 Exp: 55.4%
    • DeepSeek V4.1 Flash 1017: —

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • DeepSeek V3.2 Exp: 57.7%
    • DeepSeek V4.1 Flash 1017: —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • DeepSeek V3.2 Exp: 25.0%
    • DeepSeek V4.1 Flash 1017: —

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • DeepSeek V3.2 Exp: 33.9%
    • DeepSeek V4.1 Flash 1017: —

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • DeepSeek V3.2 Exp: 44.7%
    • DeepSeek V4.1 Flash 1017: 55.3%Top

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
17.9 tok/s
Faster than 9% of 261 models
Time to first token
3.7 s
Quicker than 47% of 261 models
Prompt cache hits
58.2%
Request success (24h)
99.8%
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.

Hourly request success · last 72 hoursLowest hour 83.3%
  • 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.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-ai/deepseek-v3.2-exp",
    "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-ai/deepseek-v3.2-exp",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

DeepSeek

Other DeepSeek models, starting with the same family.