Alibaba releases Qwen3.8-Flash-Next, targeting "ultimate cost efficiency"
decoder ·
Alibaba's Qwen team has introduced Qwen3.8-Flash-Next, a multimodal mixture-of-experts model that serves as an architecture preview for Qwen4. The model has 125 billion total parameters but activates only 6 billion per token, along with a novel 51-billion-parameter N-gram embedding layer that runs in system RAM rather than GPU memory. It natively supports a 262,144-token context window, extendable to one million tokens via YaRN, and its production version is priced at $0.16 per million input tokens and $0.47 per million output tokens. According to the team, it delivers better results than Qwen3.7-Plus at about one-ninth the training cost, with the largest gains in coding and office tasks. Published benchmarks show it outperforming DeepSeek-V4-Flash and Claude Opus 4.6 on most tasks, particularly agentic coding and productivity benchmarks, though it trails on Humanity's Last Exam. The pricing is roughly twelve times cheaper than Alibaba's flagship Qwen3.8-Max, intensifying competitive pressure on OpenAI and Anthropic.
decoder