# Zhipu AI (open.bigmodel.cn) Brand Voice

> Professional, high-performance, and utility-driven, balancing technical authority with commercial accessibility.

## Positioning
Zhipu AI is a provider of next-generation foundation models (GLM series) and developer tools. It serves enterprises and developers by offering a full-stack AI ecosystem ranging from raw API access to specialized agents and model fine-tuning.

## Voice principles
- **Technically Authoritative:** Uses precise industry terminology (multimodal, 1M context, SOTA) to establish credibility as a research-led organization.
- **Efficiency-Oriented:** Emphasizes speed and ease of integration with phrases like "one-click," "seamless," and "real-time."
- **Transparent:** Provides clear, granular pricing and technical specifications without marketing fluff or vague promises.
- **Bilingual/Global:** Maintains a professional standard in both Chinese and English, ensuring technical terms remain consistent across languages.

## Tone by context
| Context | Tone |
|---|---|
| **Marketing/Hero** | Visionary and confident. Focuses on "flagship" capabilities and "next-generation" breakthroughs. |
| **Product/Model Descriptions** | Descriptive and functional. Highlights specific use cases like "long-horizon tasks" or "complex layouts." |
| **Pricing/Documentation** | Neutral and precise. Uses structured data, clear units (per M Tokens), and logical categorization. |
| **Developer Support** | Helpful and structured. Organized by specific technical needs like "API calls" or "financial issues." |

## Lexicon
- **Use:** Flagship (旗舰), Foundation model (基座模型), Multimodal (多模态), Seamless integration (无缝融入), One-click (一键), High-performance (高性能), Long-horizon (长任务).
- **Avoid:** "Magic," "Easy" (use "Simple" or "Flexible" instead), "Cheap" (use "Cost-effective" or "Special offer"), overly emotional adjectives.

## Messaging do's and don'ts
- **Do:** Highlight the "self-developed" (自研) nature of the technology to build trust.
- **Do:** Use concrete numbers (1M context, ¥0.04/GB) to ground claims in reality.
- **Do:** Connect technical features to business value (e.g., "hitting the real needs of thousands of industries").
- **Don't:** Use hyperbolic language that isn't backed by a model version or technical report.
- **Don't:** Obscure the pricing structure; keep the "pay-as-you-go" and "flexible" nature front and center.
- **Don't:** Refer to the AI in overly personified terms; focus on it as a "tool," "agent," or "model."

## Evidence
- "Zhipu’s new flagship foundation model"
- "真正可用的 1M 上下文" (Truly usable 1M context)
- "一键接入：API 直连生产环境，无缝融入业务全流程" (One-click access: API direct to production, seamless integration)
- "按量计费，简单灵活" (Pay-as-you-go, simple and flexible)
- "The fully self-developed fourth generation"