GLM-4.7: Advancing the Coding Capability
5 months ago
- #Benchmark
- #AI
- #Coding
- GLM-4.7 introduces new features including enhanced chat, creative writing, and role-play capabilities.
- Performance benchmarks compare GLM-4.7 with models like GPT-5, Claude Sonnet 4.5, and Gemini 3.0 Pro across reasoning, coding, and agent benchmarks.
- GLM-4.7 improves Interleaved Thinking and introduces Preserved Thinking and Turn-level Thinking for complex tasks.
- The model is available via the Z.ai API platform, OpenRouter, and coding agents like Claude Code and Kilo Code.
- GLM Coding Plan subscribers get automatic upgrades to GLM-4.7, while new users benefit from cost-effective access.
- Model weights are available on HuggingFace and ModelScope, with support for vLLM and SGLang frameworks.
- Recommended settings vary by task type, with specific configurations for Terminal Bench, SWE-bench, and τ²-Bench.