Qwen 3.8 on consumer laptops matches cloud AIs for coding, claim developers
Developers report that Alibaba's open-source Qwen 3.8 models, running on consumer hardware, achieve coding quality comparable to frontier models like Anthropic's Claude Opus. Perplexity has followed by launching a local version of its agentic Computer platform, called Portable Computer, powered by Qwen 3.8. The developments could reduce dependence on cloud-based AI subscriptions.
Bottom line — Qwen 3.8 27B scored 34 on Artificial Analysis vs Claude Opus 4.6's 32 at high reasoning effort, per one developer's benchmark.
Go deeper (6)
- A developer on DEV Community running Qwen 3.8 27B on an AMD Ryzen AI Max+ 395 laptop with 128GB unified memory reported it completed complex Rust coding tasks with minimal intervention, per the developer's blog post.
- Perplexity's Portable Computer, initially limited to Nvidia's DGX Spark (around €4,100) and Linux machines with Nvidia RTX GPUs, runs Qwen 3.8 27B or Perplexity's post-trained PPLX 27B entirely on-device, according to Gizmodo.
- AMD has launched a new 'Agent Computer' category built around its Ryzen AI Max+ processors, featuring up to 128GB unified memory and 200B parameter local model support, per AMD's product page.
- The local Qwen 3.8 Flash Next variant achieved 45% fewer tokens and 3.8x faster task completion than the 27B version on five coding tasks, with equal correctness, the developer's tests showed.
- Apple's Mac Mini with M4 chip (approx. €830) was previously popular for local AI agents, but Perplexity's DGX Spark solution costs about five times that, per Gizmodo.
- Anthropic's Claude Computer Use, by contrast, remains cloud-based and gained a $30 billion Series G funding round at a $380 billion valuation, per Tech Insider.