Popular repositories Loading

1. swe-bench-88-90 swe-bench-88-90 Public

   SWE-bench Verified 90.0% / 88.0% with Qwen3.6-27B-FP8 (open-weight, 27B) + claude-cli + custom proxy. Single-machine consumer-GPU stack. Two-tier disclosure, full anti-cheat evidence.

   Shell 4

2. CogVLM2 CogVLM2 Public

   Forked from zai-org/CogVLM2

   GPT4V-level open-source multi-modal model based on Llama3-8B

   Python

3. experiments experiments Public

   Forked from SWE-bench/experiments

   Open sourced predictions, execution logs, trajectories, and results from model inference + evaluation runs on the SWE-bench task.

   Shell

4. swe-bench-submission swe-bench-submission Public

   Qwen3-Coder-Next + mini-swe-agent: SWE-bench Verified submission (342/500, 68.4%)