DeepSeek
DeepSeek: DeepSeek V4 Flash
deepseek/deepseek-v4-flash
36,5
Quality Index
115t/s
Output Speed
$0.28
Preis Output / 1M
text
Specs
| Quality Index Artificial Analysis Intelligence Index | 36,5 / 100 |
|---|---|
| Output Speed Tokens pro Sekunde | 114,8 t/s |
| Latenz (TTFT) Time-to-First-Token | 1,00 s |
| Preis Input USD je 1 Mio. Input-Tokens | $0.14 / 1M |
| Preis Output USD je 1 Mio. Output-Tokens | $0.28 / 1M |
| Context Window Maximale Tokens pro Anfrage | — |
| Release Veröffentlichungsdatum | Apr 2026 |
Über DeepSeek: DeepSeek V4 Flash
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...
Direkte Vergleiche
Wie schlägt sich DeepSeek: DeepSeek V4 Flash gegen die Top-Konkurrenz?
DeepSeek: DeepSeek V4 Flash vs. Anthropic: Claude Fable 5
Vergleich öffnen →
DeepSeek: DeepSeek V4 Flash vs. Anthropic: Claude Opus 4.8
Vergleich öffnen →
DeepSeek: DeepSeek V4 Flash vs. Google: Gemini 3.5 Flash
Vergleich öffnen →
DeepSeek: DeepSeek V4 Flash vs. MiniMax: MiniMax M3
Vergleich öffnen →