Model serving Cost Optimization Benchmarks DeepSeek V4 Flash

DSpark

Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generation

DSpark is DeepSeek's speculative-decoding framework/draft-model algorithm for large language model inference. It combines semi-autoregressive parallel block drafting with confidence-scheduled verification to speed token generation without changing the target model, and is released via DeepSpec configs/checkpoints plus DSpark-attached DeepSeek and Qwen/Gemma checkpoints.

Recent stories

1 linked story

releaseSECONDARY2026-08-03

DeepSeek V4 Flash adds Baseten and Together AI serving with 1M-token context

Baseten and Together AI added DeepSeek V4 Flash with a 1M-token context window, reasoning-effort controls, and DSpark decoding. ValsAI ranked it the cheapest model above 60 on its index, and Nous promoted a short 90% discount.