Dashboard › LLM › DeepSeek V4 Flash

LLM Model Detail

DeepSeek V4 Flash

useful $0.14 / $0.28 per M tokens

Provider: DeepSeek (via OpenRouter) · Type: Paid · Bedste til: Coding, budget tasks, high-volume routing

usefulStatus

$0.14Input/M

$0.28Output/M

2026-05-26Sidst opdateret

✅ Styrker

Best price/performance ratio for coding (75.8% SWE-bench at $0.14/M)
Fast inference speed
Excellent code generation
Great for budget-sensitive workloads

⚠️ Svagheder

Less strong on pure reasoning vs MiMo/Qwen
Occasional hallucinations on niche topics

Brug når

🎯 Use Cases

Code generation and refactoring
Budget-sensitive tasks
Bulk processing and high-volume routing
Hermes router fallback for coding queries

📝 Provider noter

Via OpenRouter. Excellent rate limits. Reliable availability with low latency. Best bang-for-buck in the paid tier.

Relaterede modeller

— Primary reasoning model