Search Articles — Sudonull

Search Results

In this project

GLM-5.1: top open-source for agentic tasks

https://sudonull.com/glm-5-1-top-open-source-for-agentic-tasks

GLM-5.1 from Z.ai sets records in SWE-Bench and NL2Repo. Learn how the model speeds up code by 6 times over 600 iterations. For developers: benchmarks, experiments, integration.

GLM-5.1 leads on SWE-Bench Pro: 58.4%

https://sudonull.com/glm-5-1-leads-on-swe-bench-pro-58-4

GLM-5.1 from Z.ai outperforms GPT-5.4 and Claude on SWE-Bench Pro. Assembled Linux desktop in 8 hours, x6 on VectorDBBench. Weights on HuggingFace. Test for long-running tasks.

Compression of GLM-5.1 on 16 GB VRAM: GQA hacks

https://sudonull.com/compression-of-glm-5-1-on-16-gb-vram-gqa-hacks

Breakdown of compressing 744B MoE model GLM-5.1 to 388 MB for T4. Fix GQA errors, config code, inference. For middle/senior dev — repeat on your hardware.

GLM-5.1 for coding: benchmark and comparison with Claude

https://sudonull.com/glm-5-1-for-coding-benchmark-and-comparison-with-claude

GLM-5.1 scores 45.3 in Z.ai coding benchmark, close to Claude Opus. Subscription limits 15x higher, access without VPN. Test in IDE and API for refactoring.

GLM-5.1 for coding: benchmarks and integration

https://sudonull.com/glm-5-1-for-coding-benchmarks-and-integration

GLM-5.1 from Zhipu AI: 94.6% of Claude Opus in coding, deploy on Huawei Ascend, prices from $3/mo. Instructions for connecting to Claude Code. Test the alternative for agentic engineering.

Trending Now