Qwen 3.8 Early Tests: Good at Agentic Coding, But Struggles with Long Loops
bindureddy · x · 2026-08-04
The author conducted preliminary tests on Qwen 3.8, noting it performs well in agentic coding but is slightly worse than Kimi 3.
He points out that current open-source models are fundamentally based on MoE (Mixture of Experts) architectures. While great at benchmark maxing, their small number of active parameters makes them unsuitable for long-running agent loops.
Related event: Alibaba's Qwen3.8-Max: Open-Source Model Nears Closed-Source Frontier(13 posts)→
More from coding & agent
- Simon Willison: code mode is the future of tool calling, if it can be sandboxed — springrod · 2026-10-16
- Dev workflow: Astra for tedious tasks, Claude for ideas not implementation — menhguin · 2026-10-16
- Agent named 'Ilya' reportedly scores 1.1pt higher on benchmarks — daniel_mac8 · 2026-10-16
- Gemini CLI fix: substring matching inlined unrequested binary files into model context — priyanshu1976 · 2026-10-16
- Building AI agents on real healthcare claims data: the domain work that makes them work — alex_verem · 2026-10-16
- OpenAI stealth-testing mysterious internal model 'Aeon' in Codex desktop app and Dots — cedric_chee · 2026-10-16