DeepSeek V4-Flash — Specifiche
Compiled by Mustafa Ihsan from the vendor’s published documentation · Last updated
| Sviluppatore | DeepSeek |
|---|---|
| Tipo | LLM (MoE) |
| Modalità | Testo → Testo |
| Parametri | 284 miliardi totali / ~13 miliardi attivi (MoE) |
| Finestra contestuale | 1 milione |
| Output massimo | 384K |
| Licenza | MIT (open) |
| Pesi aperti | Sì |
| Pubblicato | 2026-04 |
| Prezzo dell’input | $0.14 /1M |
| Prezzo dell’output | $0.28 /1M |
| Provider API | DeepSeek, OpenRouter |
Esegui localmente
| VRAM (4-bit) | ~140 GB |
|---|---|
| GPU minima | 2× H100 80 GB (4-bit) |
What is DeepSeek V4-Flash?
DeepSeek V4-Flash is the lighter member of the DeepSeek V4 family — 284B total parameters
with roughly 13B active per token, a 1M-token context window, and open MIT weights. At $0.14
in / $0.28 out per million tokens it is priced for high-volume use, and it is one of the
strongest capability-per-dollar options anywhere in the market.
The number worth internalising is the blended rate of about $0.17 per million tokens
against a frontier model’s $10. That is a ~60× spread for a model that still scores in the
respectable middle of the intelligence rankings, which makes V4-Flash the obvious candidate
for any workload where volume matters more than peak reasoning: bulk classification,
document processing, first-pass summarisation, synthetic data generation, and the retrieval
layer of a RAG pipeline. The 1M context at that price is close to unmatched. Self-hosting is
possible but not casual — about 140 GB of VRAM at 4-bit, so two H100 80GBs — which for most
teams means the API is the practical route and the open weights are insurance rather than a
deployment plan.
DeepSeek V4-Flash pricing: API cost per 1M tokens
| Input (per ogni milione di token) | $0.140 |
|---|---|
| Output (per ogni milione di token) | $0.280 |
| Rapporto output/input | 2× |
| Combinato (4:1 in:out) | $0.168 per 1 milione di token |
What DeepSeek V4-Flash costs per month
Spesa mensile reale con un mix input-to-output 4:1 — il rapporto effettivamente prodotto da un tipico carico di lavoro basato su chat o RAG.
| Carico di lavoro | Token/mese | Costo/mese |
|---|---|---|
| Progetto secondario | 1 milione in ingresso / 0,25 milioni in uscita | $0.21 |
| Piccolo team | 20 milioni in ingresso / 5 milioni in uscita | $4.20 |
| Produzione | 200 milioni in ingresso / 50 milioni in uscita | $42 |
Calcola i tuoi numeri personalizzati nel Calcolatore dei costi delle API per l'IA.
Cheaper alternatives to DeepSeek V4-Flash
| Modello | Costo combinato ($/1 milione) | Risparmi |
|---|---|---|
| Mistral 7B aperta | $0.0220 | 87% cheaper |
| Llama 3.1 8B aperta | $0.0220 | 87% cheaper |
| Mistral NeMo 12B aperta | $0.0240 | 86% cheaper |
Eseguirlo in autonomia o pagare l’API?
DeepSeek V4-Flash is open-weight, so you can run it yourself. It needs ~140 GB of VRAM at 4-bit (2× H100 80GB (4-bit)). Self-hosting only beats the API once your volume is high enough to keep that hardware busy — the calcolatore self-hosting vs API calcola il punto di pareggio in base al tuo volume di token.
Domande frequenti
How much does DeepSeek V4-Flash cost per 1M tokens?
DeepSeek V4-Flash costs $0.140 per 1M input tokens and $0.280 per 1M output tokens. At a typical 4:1 input-to-output mix that blends to about $0.168 per 1M tokens.
How much does DeepSeek V4-Flash cost per month?
A small-team workload of 20M input and 5M output tokens a month costs about $4.20 on DeepSeek V4-Flash. A side project (1M in / 0.25M out) costs roughly $0.21.
What is a cheaper alternative to DeepSeek V4-Flash?
Mistral 7B is the strongest cheaper option in our database at $0.0220 per 1M blended — about 87% less than DeepSeek V4-Flash. It is also open-weight, so self-hosting is an option.
Can I run DeepSeek V4-Flash locally?
Yes. DeepSeek V4-Flash is open-weight and needs about ~140 GB of VRAM at 4-bit quantisation (2× H100 80GB (4-bit)).
Why does DeepSeek V4-Flash charge more for output than input?
Output tokens are generated one at a time and cannot be batched the way a prompt can, so they cost the provider more to serve. DeepSeek V4-Flash charges 2× more for output, which is why prompt-heavy workloads are far cheaper to run than generation-heavy ones.
See every DeepSeek model priced side by side: Prezzi dell'API DeepSeek.
I prezzi indicati sono le tariffe ufficiali pubblicate per l'API principale del modello e vengono aggiornati man mano che i fornitori li modificano. Sconti per volumi elevati, elaborazione batch e input memorizzati nella cache non sono inclusi. Confronta tutti i modelli fianco a fianco nel Database di modelli IA o il Classifica LLM.