GLM-4.6 Available on Novita AI: Zai-org’s New Generation Flagship Model with 200K Context Window

GLM-4.6 is now available on the Novita AI platform, bringing Zai-org’s new generation flagship model with major improvements in context length, coding performance, and agentic capabilities. Featuring a 355B-parameter MoE (Mixture of Experts) architecture and achieving state-of-the-art performance among open source models, GLM-4.6 represents a significant advancement in AI capabilities.

This latest release doubles the context window from 128K to 200K tokens while achieving near-equivalent performance to Claude Sonnet 4 in real-world coding tasks. Whether you’re building AI agents, developing complex applications, or creating automation solutions, GLM-4.6 delivers the capabilities you need through Novita AI’s developer-friendly infrastructure.

Current pricing on Novita AI: 204,800 Context, $0.6/1M input tokens, $2.2/1M output token

What is GLM-4.6?

GLM-4.6 is Zhipu AI’s new generation flagship model that brings significant improvements over GLM-4.5, achieving state-of-the-art performance among open source models. Built with a 355B-parameter MoE architecture, it’s specifically designed to excel in agentic tasks, coding applications, and complex reasoning scenarios.

Expanded Context Window: GLM-4.6 introduces a 200K token context window (up from 128K in GLM-4.5), allowing it to handle more complex conversations and process larger codebases. This expansion lets developers work with extensive documentation, analyze longer code files, and maintain context across sophisticated agent workflows.

Superior Coding Performance: GLM-4.6 shows substantial improvements across multiple benchmarks and exceptional real-world performance in popular coding assistants like Claude Code, Cline, Roo Code, and Kilo Code. The model excels at generating visually polished front-end pages and handling complex development tasks with greater accuracy.

Enhanced Reasoning Capabilities: The model’s reasoning has been strengthened through support for tool use during inference, leading to better performance in problem-solving scenarios. GLM-4.6 integrates more effectively within agent frameworks, making it ideal for building AI-powered automation systems that require multi-step reasoning and external tool integration.

Refined Writing Quality: GLM-4.6 produces writing that better aligns with human preferences in style and readability, performing more naturally in role-playing scenarios and content generation tasks.

Performance Benchmarks

GLM-4.6 demonstrates strong performance across comprehensive evaluations covering agents, reasoning, and coding capabilities.

Public Benchmark Results

Evaluated across eight public benchmarks, GLM-4.6 shows clear improvements over GLM-4.5 and achieves state-of-the-art performance among open source models. It holds competitive performance against leading models like DeepSeek-V3.2-Exp and Claude Sonnet 4, though it still falls behind Claude Sonnet 4.5 in pure coding ability.

Real-World Performance (CC-Bench)

In the extended CC-Bench evaluation, human evaluators used GLM-4.6 inside isolated Docker containers to complete multi-turn tasks across front-end development, tool building, data analysis, testing, and algorithm implementation.

The results show GLM-4.6 reaches near-equivalent performance to Claude Sonnet 4, achieving a 48.6% win rate while clearly outperforming other open-source models.

Token Efficiency

GLM-4.6 completes tasks with approximately 15% fewer tokens than GLM-4.5, resulting in faster response times, lower computational costs, and maintained or improved output quality.

Getting Started with GLM-4.6 on Novita AI Platform

Novita AI offers multiple ways to access GLM-4.6, designed for different skill levels and use cases.

Use the Playground (No Coding Required)

Sign up and start experimenting with GLM-4.6 in seconds through an interactive interface. Test prompts, see outputs in real-time with the full 200K context window, and compare GLM-4.6 with other leading models. Perfect for prototyping and understanding what the model can do before building full implementations.

Integrate via API (For Developers)

Connect GLM-4.6 to your applications using Novita AI’s unified REST API.

Direct API Integration (Python Example)

 from openai import OpenAI
  
client = OpenAI(
    base_url="https://api.novita.ai/openai",
    api_key="",
)

model = "zai-org/glm-4.6"
stream = True # or False
max_tokens = 49152
system_content = "Be a helpful assistant"
temperature = 1
top_p = 1
min_p = 0
top_k = 50
presence_penalty = 0
frequency_penalty = 0
repetition_penalty = 1
response_format = { "type": "text" }

chat_completion_res = client.chat.completions.create(
    model=model,
    messages=[
        {
            "role": "system",
            "content": system_content,
        },
        {
            "role": "user",
            "content": "Hi there!",
        }
    ],
    stream=stream,
    max_tokens=max_tokens,
    temperature=temperature,
    top_p=top_p,
    presence_penalty=presence_penalty,
    frequency_penalty=frequency_penalty,
    response_format=response_format,
    extra_body={
      "top_k": top_k,
      "repetition_penalty": repetition_penalty,
      "min_p": min_p
    }
  )

if stream:
    for chunk in chat_completion_res:
        print(chunk.choices[0].delta.content or "", end="")
else:
    print(chat_completion_res.choices[0].message.content)
  

Multi-Agent Workflows with OpenAI Agents SDK

Build sophisticated multi-agent systems with plug-and-play integration, support for handoffs, routing, and tool integration with the full 200K context window.

Connect with Third-Party Platforms

Coding Agents: Integrate with popular coding assistants like Claude Code, Cursor, Codex, Trae, Qwen Code, and Cline through OpenAI-compatible APIs and Anthropic-compatible APIs.

Orchestration Frameworks: Connect with LangChain, Dify, CrewAI, and Langflow using official connectors.

Hugging Face: Novita AI is an official inference provider for Hugging Face, ensuring broad ecosystem compatibility.

Conclusion

GLM-4.6 on Novita AI delivers Zhipu AI’s new generation flagship model with a 355B-parameter MoE architecture and 200K context window, achieving state-of-the-art performance among open source models. With near-equivalent performance to Claude Sonnet 4 (48.6% win rate) and 15% better token efficiency than GLM-4.5, GLM-4.6 represents a significant leap forward in accessible AI capabilities.

Start exploring GLM-4.6 today through Novita AI’s playground, API, or third-party integrations to enhance your development workflow with exceptional coding assistance, refined writing, and powerful reasoning capabilities.

Novita AI is an AI cloud platform that offers developers an easy way to deploy AI models using our simple API, while also providing affordable and reliable GPU cloud for building and scaling.