- 🔒 Runs in a Trusted Execution Environment with hardware attestation.
- 🔧 Z.ai's flagship tuned for agentic coding and terminal workflows.
- 🧠 Adds Preserved Thinking and Turn-level Thinking atop interleaved reasoning.
- 👁️ Improved frontend/UI generation with cleaner layouts and styling.
- 📏 Catalog lists a 128K-token context window.
- 🏢 Mixture-of-Experts design, released under the permissive MIT license.
- 🌐 Supports web search and multi-step agent execution.
- 🆕 Works natively with Claude Code, Cline, Roo Code, Kilo Code.
Z.ai, formally Knowledge Atlas Technology Joint Stock Co., Ltd., is a Chinese technology company specializing in artificial intelligence. Previously known internationally as Zhipu AI, the company rebranded to Z.ai in 2025. Its core focus is the GLM family of large language…
Explore 17 more models by Z.ai →GLM 4.7 is Z.ai's flagship text model, offered here in a confidential-computing configuration: per the catalog and provider description, it runs inside a Trusted Execution Environment (TEE) so that hardware attestation evidence can be independently verified. Z.ai (formerly Zhipu AI) ships its GLM family under the MIT license, and GLM 4.7 is positioned for "task delivery" — an end-to-end loop from requirement understanding to executable code across multiple programming languages.
Relative to its same-family predecessors such as GLM 4.6, GLM 4.7 concentrates on two upgrades: stronger programming capability and more stable multi-step reasoning. Its headline change is a three-tier thinking design that extends the Interleaved Thinking introduced in GLM-4.5 by adding Preserved Thinking and Turn-level Thinking, so reasoning persists across actions and conversational turns instead of restarting each round. This is aimed at the "context collapse" problem in long coding sessions. Z.ai's documentation also highlights improved visual-code and UI understanding for layout, color harmony, and component styling.
On specifications, the catalog lists a 128K-token context window, and the model uses a Mixture-of-Experts architecture.
Within this confidential-computing line, GLM 4.7 precedes the newer GLM 5.2. For users who need verifiable execution privacy alongside agentic coding, the TEE build pairs GLM 4.7's capabilities with attestable hardware isolation.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Routing | Input $/M | Cached $/M | Output $/M | Categories | API |
|---|---|---|---|---|---|---|---|
| antseed-neon-puma-944e 0x6650…944e | 0.00 | gated | $0.3751 | $0.3751 | $1.4152 | chat,coding,math | openai-chat-completions |
| Leftermute 0x388b…5389 | 0.00 | gated | $0.2222 | $0.2222 | $0.8383 | chat,coding,json,tools | openai-chat-completions |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = on-chain trust (0-100). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on AntSeed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.