Unknown · v5.2 · 34× · vu le 25 juil. 2026
80
Momentum
GLM-5.2 is the flagship language model from Z.ai (Zhipu AI) for long-horizon, agentic tasks and coding. It's a Mixture-of-Experts model with approximately 744-753 billion total parameters (roughly 40 billion active per Token) and a true 1-million-Token context window. The model is distributed under an MIT Open Source license without regional restrictions and ranks as the most powerful open model on the Artificial Analysis Intelligence Index and across several coding Benchmarks. GLM-5.2 supports two Reasoning modes (High and Max) and was first released on June 13, 2026 to paying users of the GLM Coding Plan, followed by API and open weights.
Historique du momentum
27.04.26.07.
Fonctionnalités
| Key Benchmark (%) | Artificial Analysis Intelligence Index: 51 points (leading open-weight model); Terminal-Bench 2.1: 81.0%; SWE-bench Pro: 62.1% |
| Context Window (Tokens) | 1,000,000 tokens (max output 131,072 tokens) |
| License | MIT license (open source, no regional restrictions) |
| Multimodality | Text-only (input and output), no image understanding |
| Platform | API via Z.ai platform and providers like DeepInfra, Fireworks, FriendliAI; weights on Hugging Face and ModelScope; local deployment via vLLM, SGLang, Transformers, KTransformers |
| Price per 1M Tokens | $1.40 per 1M input tokens / $4.40 per 1M output tokens (Z.ai API) |
| Release Date | June 13, 2026 (Coding Plan); MIT open weights around June 16-17, 2026 |