Back to models
DeepSeek V4 Flash
Fast, lightweight DeepSeek V4 model with very long context.
Context
1,000,000
≈ 1M
Max output
384K
384,000
Price ($/1M tokens)
—
Specs
Input
Text
Output
Text
Reasoning
Yes
Tool call
Yes
Structured
Yes
Weights
Open
Released
2026-04-24
Free on relays
2
