Generation Requires: https://github.com/vllm project/llm compressor/pull/1788 Evaluation The model was evaluated on HumanEval and HumanEval+ benchmark with the Neural Magic fork of the EvalPlus implementation of HumanEval+ and the vLLM engine, using the following commands: Metric Qwen/Qwen3 Coder 30B A3B Instruct nm testing/Qwen3 Coder 30B A3B Instruct W4A16 awq : : : : HumanEval pass@1 93.0 93.7 HumanEval pass@10 93.9 94.5 HumanEval+ pass@1 88.7 89.3 HumanEval+ pass@10 89.8 90.2 Average Score 91.35 91.93
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy