Compare two models
Select any two models. Input, output, cached, and context rarely move together — the cheaper model on one line isn't always the cheaper one on the next.
Description
A sparse mixture-of-experts model from DeepSeek built on its Causal Encoder-Decoder architecture, activating 8 billion parameters on input and 16 billion on output from a 284-billion-parameter pool, with a 1-million-token context window.
OpenAI's frontier model for complex professional workloads. 1M token context, text + image input, native computer use.
I/O price /1M
$0.112/$0.336
$5.00/$30.00
Input /1M
$0.112
$5.00
Output /1M
$0.336
$30.00
Cached in /1M
$0.0034
$0.5
Context
1.05M
1M
Δ input since launch
—
—
Δ output since launch
—
—
Released
Sep 2026
Apr 2026
Status
—
—
Provider
DeepSeek
OpenAI