Back to models
DeepSeek logo

DeepSeek

deepseek-chat (DeepSeek-V3.2)

DeepSeek’s non-thinking API SKU mapped to the DeepSeek-V3.2 model version.

Overall score
86
budgetcodingapi-sku
Context window
128K tokens
Speed
Fast
Input pricing
$0.28 / 1M input tokens (cache miss)
Output pricing
$0.42 / 1M output tokens

Score breakdown

Capability84
Use-case fit86
Cost efficiency96
Speed92
Reliability82
Agent readiness84
Ecosystem80

Scores combine benchmark signals, product experience, and editorial weighting. Use them as a practical guide, not an absolute truth claim.

Best for

CodingAgent automation

Works with

DeepSeek APIrouting layersbudget coding flows

Modalities

text

Sources & trust

Officially verified core fields
Official linkSummaryDescriptionContext windowMax outputTool supportPricingPricing page

Editorial fields such as shortlist guidance, strengths, caveats, and scoring remain clearly separated from official provider data.

Last verified: Apr 9, 2026

Strengths

  • Very aggressive API pricing for practical text workloads.
  • Useful as a cheaper mainline or fallback model in routed stacks.

Things to watch

  • API SKU naming differs from the broader model-family branding.
  • Some enterprises will still prefer vendors with broader compliance posture.

Best for