Skip to content
AI IntelligenceAug 3, 2026AI Intelligence
Article

DeepSeek officially released and open-sourced V4-Flash, a 304-billion parameter lightweight model that outperforms its V4-Pro preview.

Priced at $0.14/$0.28 per million tokens, it tops VulcanBench benchmarks and matches Claude Opus-4.8 performance, intensifying competition in the open-weight frontier tier.

Data Cube AI EditorialSource: Pandaily
01

Source Brief

DeepSeek officially released and open-sourced V4-Flash, a 304-billion parameter lightweight model that outperforms its V4-Pro preview. Priced at $0.14/$0.28 per million tokens, it tops VulcanBench benchmarks and matches Claude Opus-4.8 performance, intensifying competition in the open-weight frontier tier.