Z.AI: GLM 5
z-ai/glm-5
Try GLM 5 for free in your browser, and add it to your app for free with Puter.js AI API.
Try it free Add to your app// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
puter.ai.chat("Explain quantum computing in simple terms", {
model: "z-ai/glm-5"
}).then(response => {
document.body.innerHTML = response.message.content;
});
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
puter.ai.chat("Explain quantum computing in simple terms", {
model: "z-ai/glm-5"
}).then(response => {
document.body.innerHTML = response.message.content;
});
</script>
</body>
</html>
Model Card
GLM-5 is Zhipu AI's (Z.ai) fifth-generation flagship open-weight foundation model with 744B total parameters (40B active) in a Mixture of Experts architecture, designed for agentic engineering, complex systems coding, and long-horizon agent tasks. It achieves state-of-the-art performance among open-weight models on coding and agentic benchmarks like SWE-bench Verified and Terminal Bench 2.0, approaching Claude Opus 4.5-level capability.
Context Window 200K
tokens
Max Output 128K
tokens
Input Cost $1
per million tokens
Output Cost $3.2
per million tokens
Input text
modalities
Tool Use Yes
Release Date Feb 12, 2026
Try GLM 5 for free
Try GLM 5 instantly in your browser.
This playground uses the Puter.js AI API — no API keys or setup required.
Benchmarks
How GLM 5 performs on standard evaluations.
| Benchmark | Score |
|---|---|
| GPQA Diamond Graduate-level science Q&A | 82.0% |
| Humanity's Last Exam Cross-domain reasoning | 29.3% |
| IFBench Instruction following | 72.3% |
| LCR Long-context reasoning | 75.7% |
| Terminal-Bench Hard Agentic terminal tasks | 43.2% |
| τ²-Bench Tool use / agents | 98.2% |
Scores sourced from Artificial Analysis.
Find other Z.AI models →
GLM 5.3 Prime
GLM 5.3 Prime is a high-speed serving tier built on Z.ai's GLM-5.3, offered through third-party inference platforms rather than as an official Z.ai product listing. It runs the same 744-billion-parameter Mixture-of-Experts weights (40 billion active) as standard GLM-5.3, on infrastructure tuned for higher throughput. Vendor reports put its output throughput at 1.5 to 2 times the standard GLM-5.3 endpoint, at roughly double the per-token price, a pattern also seen with GLM-5.2's own "Fast" tier. No independent benchmark of the Prime endpoint exists, but since the weights are unchanged, it should inherit GLM-5.3's reported gains over GLM-5.2, including a 50% jump on Z.ai's Code Bench and gains on Terminal-Bench 3.0 and DeepSWE v1.1. It supports a 1,000,000-token context window, 128,000 max output tokens, three reasoning-effort levels, and function calling. Best suited for agent loops and coding assistants that chain many model calls and are latency-sensitive, where the throughput gain offsets cost.
ChatGLM 5.3 FlashX
GLM 5.3 FlashX is Z.ai's high-speed serving tier for GLM-5.3-Flash, using the same 320-billion-parameter Mixture-of-Experts weights (18 billion active) rather than a separate model. Z.ai runs it on faster inference infrastructure, reporting speeds up to 200 tokens per second. Because the weights are unchanged, it inherits GLM-5.3-Flash's benchmark results, including 63.4 on DeepSWE v1.1 and 48.8 on AutomationBench, both ahead of GLM-5.2. It keeps the same 1,000,000-token context window, 128,000-token max output, and function calling support. Pricing is $0.37 per million input tokens and $1.25 per million output tokens, about 2.5 times GLM-5.3-Flash's rate. It fits agent loops and coding assistants that call the model repeatedly and need lower per-call latency, where the added throughput offsets the higher per-token price.
ChatGLM 5.3 Flash
GLM 5.3 Flash is a mixture-of-experts model from Z.ai (Zhipu AI), the first natively multimodal release in the GLM-5 family, accepting text and image input. It has 320 billion total parameters with 18 billion active per token, and a hybrid sparse-and-linear attention architecture that Z.ai says cuts attention compute about 3x and KV cache size about 4.4x compared with GLM-5.3. Through this API it offers a 1,000,000 token context window and up to 131,100 output tokens, priced at $0.08 per million input tokens and $0.25 per million output tokens. It supports function calling, and thinking is always on and cannot be disabled. GLM 5.3 Flash beats GLM-5.2 on Z.ai's reported benchmarks (DeepSWE v1.1, 63.4 vs 46.2; AutomationBench, 48.8 vs 26.2) and approaches Claude Opus 4.8 on coding and agentic tasks. Model weights are released under the MIT license.
Frequently Asked Questions
You can access GLM 5 by Z.AI through Puter.js AI API. Include the library in your web app or Node.js project and start making calls with just a few lines of JavaScript — no backend and no configuration required. You can also use it with Python or cURL via Puter's OpenAI-compatible API.
GLM 5 is free to try with a Puter account. Every account includes a free AI allowance, and you can chat with it in the playground on this page. You can upgrade your account anytime for a larger allowance.
GLM 5 is free to integrate using the Puter.js AI API. With the User-Pays Model, you can add AI to your app for $0, since users cover their own AI usage through their Puter account.
| Price per 1M tokens | |
|---|---|
| Input | $1 |
| Output | $3.2 |
GLM 5 was created by Z.AI and released on Feb 12, 2026.
GLM 5 supports a context window of 200K tokens. For reference, that is roughly equivalent to 400 pages of text.
GLM 5 can generate up to 128K tokens in a single response.
GLM 5 accepts the following input types: text. It produces: text.
Yes, GLM 5 supports tool use (function calling), allowing it to interact with external tools, APIs, and data sources as part of its response flow.
GLM 5 scores 27.9 on the Artificial Analysis Intelligence Index, outperforming 82% of tracked models.
Yes — the GLM 5 API works with any JavaScript framework, Node.js, or plain HTML through Puter.js. Just include the library and start building. See the documentation for more details.
Add GLM 5 to your app for free
Developers can integrate GLM 5 for free using the Puter.js AI API.
With the User-Pays Model, each user covers their own AI usage instead of the developer.