Models / DeepSeek / DeepSeek V3 0324

DeepSeek · deepseek

DeepSeek V3 0324

March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding

$0.24 in · $0.90 out

per 1M tokens · provider rate closest to the median. No first-party DeepSeek price was listed. The reference price is Deep Infra's rate, the provider closest to the median.

Provider input prices run from $0.20 to $0.50 per 1M tokens. Output prices run from $0.77 to $1.50.

Context
163,840
Max output
163,840
Cache read
$0.14
Cache write
—
Reasoning
No
Tool calls
Yes
Structured output
Unknown
Vision
No
Open weights
Yes
License
—
Released
Mar 24, 2025
Updated in source
Mar 24, 2025
Knowledge
—
Providers
5

Model weights weights

What this model costs per month

Prices are USD per 1M tokens. A month is 30 days. Cache hits use the cache-read rate when the model has one; otherwise the cache field is ignored.

monthly = requests/day × 30 × (input tokens × (1 − cache hit) × input price + input tokens × cache hit × cache-read price + output tokens × output price) ÷ 1,000,000

Providers

Base rates from Models.dev. One row per provider.
ProviderInputOutputCache readCache writeStatus
NanoGPT$0.20$0.77$0.14——
Deep Infrareference$0.24$0.90$0.14—deprecated
Eden AI$0.24$0.90$0.14——
Hugging Face$0.27$1.12———
Crusoe$0.50$1.50$0.25——

Jobs this model qualifies for