General Language Model - 4.6
General Language Model - 4.6 is a large-scale foundation model developed by Zhipu AI. With approximately 3.5 trillion parameters, it is a chat-specialized model that supports a long context window of up to 200K.
Parameters
3550.0B
Context Window
200K
License
MIT
Release Date
2025-09-30
API Pricing
API pricing for this model is not yet available
Strengths
- ・Massive 3.5 trillion parameters
- ・200K long context window
- ・High openness via MIT license
Weaknesses
- ・Enormous file size over 700GB
- ・Requires extreme computational resources
- ・Difficult to run in lightweight environments
Use Cases
- ・Building advanced conversational AI services
- ・Analyzing ultra-long documents
- ・Constructing large-scale knowledge bases
Deep Analysis
Architecture
MoE (357B active)
Expanded from GLM-4.5
Context Window
200K tokens
Up from 128K in GLM-4.5
Maximum Output
128K tokens
Release Date
September 2025
License
Open weights
Key Improvement
Coding + Agents
Performance on par with Claude Sonnet 4/4.6
Strengths
- ・200K context window (up from 128K)
- ・Performance on par with Claude Sonnet 4/4.6
- ・Strong real-world coding in Claude Code environment
- ・30% more token-efficient than GLM-4.5
- ・Enhanced reasoning with tool use during inference
- ・Better writing aligned with human preferences
Weaknesses
- ・Chinese model with English as secondary
- ・Large model requires significant compute
- ・Competes with many strong alternatives
Competitor Comparison
| Model | Arena | SWE | GPQA | Price |
|---|---|---|---|---|
| Claude Sonnet 4.6 | - | - | - | Higher |
| GLM-4.5 355B | - | - | - | Lower |
| DeepSeek V3.2 | - | - | - | Comparable |
GLM-4.6 is Zhipu AI's enhanced language model achieving performance on par with Claude Sonnet 4/4.6 across 8 authoritative benchmarks. Released September 2025, it expands context to 200K tokens and significantly improves coding, reasoning, and agentic capabilities.
Sources
Analysis generated: 2026-05-24