DeepSeek V4 Flash 0423

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...

De onderstaande beschrijving is overgenomen van OpenRouter en is in het Engels.

Uitgebracht

Context window

1M tokens

Max. output

944K tokens

Input

tekst

Output

tekst

Reasoning

Ja

Prijshistorie

• input $0.14 → $0.091 -35% · output $0.28 → $0.182 -35%
• input $0.047 → $0.14 +196% · output $0.095 → $0.28 +196%
• input $0.045 → $0.047 +5% · output $0.45 → $0.095 -79%
• input $0.047 → $0.045 -5% · output $0.45 → $0.45 0%
• input $0.05 → $0.047 -5% · output $0.45 → $0.45 0%
• input $0.14 → $0.05 -64% · output $0.28 → $0.45 +61%
• input $0.091 → $0.14 +54% · output $0.182 → $0.28 +54%
• input $0.05 → $0.091 +82% · output $0.45 → $0.182 -60%
• input $0.091 → $0.05 -45% · output $0.182 → $0.45 +147%

Aanbieders op OpenRouter

17 endpoints bieden dit model aan. Uptime is het aandeel geslaagde verzoeken via OpenRouter; prijzen per 1M tokens.