deepseek-flash
DeepSeek's fast, inexpensive model for high-throughput and latency-sensitive workloads. Open weights under MIT, it is a budget-friendly companion to the v4-pro flagship.
At a glance
Released
Aug 20, 2026
Provider
DeepSeek
Context window
256,000 tokens
Max output
16,000 tokens
Modalities
Text → text
Knowledge cutoff
May 2026
API model ID
deepseek-flashLicense
Open weights (MIT)
Pricing
Standard
Input$0.27 / 1M
Output$1.10 / 1M
* Rates are from a third-party aggregator — verify before relying on them. A typical 1,500-token prompt with a 650-token answer costs ≈ $0.001 on the standard tier. Run your own prompt through the calculator →
What's new vs the previous model
Compared withdeepseek-v3-flash (February 2026)
- Lower latency and higher throughput
- Improved instruction following for its size
- Better short-context reasoning
- Smaller context and output than v4-pro
- Text-only, no image input
Best use cases
High-volume tasks
Runs classification and routing cheaply at scale.
Self-hosted serving
Open weights allow local or private deployment.
Quick chat
Powers simple assistant flows with low latency.
Strengths & limitations
Strengths
- Very cheap
- Low latency
- Open weights
- High throughput
Limitations
- Limited deep reasoning
- Text-only
How to test it
- No-code: open chat.deepseek.com and pick deepseek-flash from the model switcher — no API key needed.
- Playground: sign in to the DeepSeek Platform and run prompts in the workbench, then copy the generated code.
- API: create a key in the DeepSeek Platform and call the API (example below).
curl https://api.deepseek.com/chat/completions \
# OpenAI-compatible endpoint
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-flash",
"messages": [{"role": "user", "content": "Explain tokenization in one paragraph."}]
}'