# Chinese Open-Weight Models Set the Coding Bar for Downloadable Systems in August

Trackers put DeepSeek V4 Pro at 80.6 percent on SWE-bench Verified and Alibaba's Apache-licensed Qwen3.8-27B at 73.0 on Terminal-Bench 2.1, both released within a day of each other.

- Published: 2026-08-22T06:18:28.650Z
- Canonical: https://polylog.news/ai/2026-08-22/chinese-open-weight-models-set-the-coding-bar-for-downloadab
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [MindStudio](https://www.mindstudio.ai/blog/deepseek-v4-launch-specs-open-weight-2026), [OrcaRouter](https://www.orcarouter.ai/blog/qwen-3-8-27b-benchmarks), [Wavect](https://wavect.io/blog/open-weight-llm-comparison-2026/)

Two downloadable releases arrived a day apart in mid-August. DeepSeek published V4 Pro on August 13, with trackers reporting 80.6 percent on SWE-bench Verified, the strongest published coding result from an open-weight model to date. Alibab…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-08-22/chinese-open-weight-models-set-the-coding-bar-for-downloadab (subscription information: https://polylog.news/pricing).