Model2h ago
Alibaba Previews Qwen4 With High-Efficiency Flash Model
The Qwen3.8-Flash-Next model activates only 6 billion of its 125 billion parameters per token. It outperforms DeepSeek-V4-Flash and Claude Opus 4.6 on coding benchmarks despite costing one-ninth as much to train. Alibaba is using this architecture to undercut OpenAI and Anthropic on pricing. Developers gain a cheaper, high-performance alternative for office automation.