GLM 4.7
Z.AI's open-weight coding and reasoning model that runs privately on Venice with tool use and zero retention.
Overview
What is GLM 4.7
GLM 4.7 is Z.AI's open-weight text model released in 2025, optimized for agentic coding, reasoning, and tool use. It supports function calling, web search, and structured output, runs under the MIT license, and is available on Venice with zero prompt retention as the default model for tool use.
Running it privately on Venice
On Venice, GLM 4.7 runs with zero retention — your prompts are not stored, profiled, or used for training. It is the default model for tool use and carries the 'Most intelligent' trait, offering function calling, reasoning, and web search without Big-Tech surveillance. You pay only for tokens used, with no subscription lock-in.
Assessment
Strengths and limitations
- Open weights under MIT license, enabling self-hosting and fine-tuning outside Venice.
- Strong agentic coding and terminal-task performance, with native support for reasoning before acting.
- Built-in tool use: function calling, web search, and structured JSON output for system integration.
- Cost-efficient per-token pricing compared to closed frontier models.
- Default tool-use model on Venice with the 'Most intelligent' trait.
- Quantized to fp4 on Venice, which may trade marginal precision for speed and cost.
- Not uncensored; it retains standard safety alignment.
- GLM 5.1 surpasses it on the latest SOTA agentic coding benchmarks.
- Text-only: no native vision or image understanding.
Capabilities
What it supports
- Tool use / function calling
- Vision (image input)
- Reasoning
- Web search
- Code-optimized
- Structured output (JSON schema)
- Audio input
- Video input
- Multiple image inputs
- Log probabilities
Specifications
Datasheet
- Maker
- Z.AI (Zhipu AI)
- Released
- 2025
- Modality
- Text
- License
- MIT
- Open weights
- Yes
- Context window
- 198K tokens
- Max output
- 16.384K tokens
- Capabilities
- Function calling, Reasoning, Web search
- Privacy on Venice
- Private — zero retention
- Available on Venice since
- Dec 2025
API
Call it from your code
Venice exposes an OpenAI-compatible API. Point your base URL at Venice and pass the model id.
curl https://api.venice.ai/api/v1/chat/completions \
-H "Authorization: Bearer $VENICE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "zai-org-glm-4.7",
"messages": [{ "role": "user", "content": "Explain quantum tunneling simply." }]
}'Pricing
What it costs on Venice
Billed per token on Venice: $0.55 per 1M input tokens and $2.65 per 1M output tokens.
New Venice accounts include a free daily allowance and 500 welcome credits — no credit card required.
Alternatives
How it compares
| Model | Best for | Price (Venice) | Context window | Open weights |
|---|---|---|---|---|
| GLM 4.7 | The balanced open-weight workhorse — strong coding, tool use, and vibe generation at a mid-tier price. | $0.55 in · $2.65 out / 1M | 198K tokens | Yes |
| DeepSeek V3.2 | Lower per-token cost and solid reasoning, but GLM 4.7 leads on agentic coding and UI tasks. | $0.33 in · $0.48 out / 1M | 160K tokens | Yes |
| Kimi K2.6 | A larger-context open rival with comparable coding strength; GLM 4.7 is cheaper and Venice's default for tool use. | $0.75 in · $3.50 out / 1M | 256K tokens | Yes |
| GLM 5.1 | Z.AI's successor flagship — upgrade only when you need the absolute latest SOTA agentic performance. | $1.10 in · $4.15 out / 1M | 200K tokens | Yes |
The balanced open-weight workhorse — strong coding, tool use, and vibe generation at a mid-tier price.
Use cases
What it is good for
- 01Agentic coding assistants and terminal-based automation.
- 02Tool-using AI agents that require web search or API orchestration.
- 03Vibe coding and UI generation with modern layout quality.
- 04High-volume chat and reasoning workloads where open weights and cost matter.
Prompting
Getting better results
Use structured JSON schema when integrating with external systems.
Enable reasoning mode for multi-step math and logic tasks.
For coding, provide project-level context and clear engineering standards to maximize agentic performance.
FAQ
Frequently asked questions
GLM 4.7 is Z.AI's open-weight text model released in 2025, optimized for agentic coding, reasoning, and tool use. It supports function calling, web search, and structured output, and is available on Venice with zero prompt retention.
On Venice, GLM 4.7 is billed at $0.55 per 1M input tokens and $2.65 per 1M output tokens, with cached input at $0.11 per 1M. There is no subscription required.
Yes. GLM 4.7 is released under the MIT license with open weights available on Hugging Face, so you can self-host or fine-tune it. Running it on Venice is pay-per-token.
Yes. It supports function calling, web search, and structured JSON output, making it well-suited for agentic workflows and coding assistants.
No. While it runs privately on Venice with zero retention, the model itself is not uncensored and retains standard safety alignment.
GLM 5.1 is Z.AI's newer flagship with stronger SOTA coding performance, but GLM 4.7 is cheaper per token and still excellent for general coding, tool use, and chat. Choose 4.7 for cost efficiency and mature stability; choose 5.1 for maximum agentic coding power.
It runs under Venice's private tier with zero retention — your prompts are not stored, profiled, or used for training. It does not currently run in a TEE or end-to-end encrypted session.
Agentic coding, terminal-based tasks, vibe coding with UI generation, tool-using agents, and complex reasoning that requires function calling or web search.
Run GLM 4.7 privately
No prompt logging. No data used for training.