GLM 4.7 vs GPT-5.6 Terra
Pricing, context window and real answers (August 2026)
GLM 4.7 (Z.ai) and GPT-5.6 Terra (OpenAI), compared as you can actually call them on FastMetal. Both are served from the same OpenAI-compatible endpoint and API key; switching is a change to the model string.
Specs and pricing
 GLM 4.7 |  GPT-5.6 Terra | |
|---|---|---|
| Provider | Z.ai | OpenAI |
| Input (per 1M tokens) | ¥107.22 | ¥357.4 |
| Output (per 1M tokens) | ¥393.14 | ¥2,144.4 |
| Typical cost (1,000 in + 500 out tokens × 1,000 calls) | ¥304 | ¥1,430 |
| Context window | 202,752 tokens | 1,050,000 tokens |
| Release date | 12/22/2025 | 7/9/2026 |
| Input modalities | text | file, image, text |
| Arena · overall | #79 · ELO 1,442 | #42 · ELO 1,466 |
| Arena · Japanese | Unranked | #19 · ELO 1,474 |
| Arena · coding | #87 · ELO 1,485 | #31 · ELO 1,519 |
Prices are FastMetal's yen rates before tax. Arena ranks are as of the latest leaderboard scrape.
Which should you pick?
- On price, GLM 4.7: ¥393.14 per 1M output tokens, 82% less than GPT-5.6 Terra.
- On the Overall arena board, GPT-5.6 Terra ranks higher (#42; GLM 4.7 is #79).
- On the Coding arena board, GPT-5.6 Terra ranks higher (#31; GLM 4.7 is #87).
- For long documents, GPT-5.6 Terra: a 1,050,000-token context window against 202,752 for GLM 4.7.
- If in doubt, try both on the same key. Switching is a change to the model string, and each is billed at its own rate.
Real answers to the same prompts
Exactly what the FastMetal gateway returned, side by side. Not benchmark scores: actual output.
Count the number of 'r's in 'strawberry'
Count the number of 'r's in 'strawberry'. Explain your reasoning step by step.
GLM 4.7
There are **3** "r"s in "strawberry".
Here is the step-by-step reasoning:
1. **Break down the word into individual letters:**
s - t - r - a - w - b - e - r - r - y
2. **Identify the position of each 'r':**
* The 3rd letter is **r**.
* The 8th letter is **r**.
* The 9th letter is **r**.
3. **Count the total occurrences:**
1 + 1 + 1 = 3.
23.3s685 output tokens
GPT-5.6 Terra
“strawberry” is spelled:
s t **r** a w b e **r** **r** y
There are **3** occurrences of the letter **r**.
3.6s71 output tokens
Debug This Error
I'm getting the following error in my Node.js application: TypeError: Cannot read properties of undefined (reading 'map') at UserList (/app/components/UserList.js:12:25) at renderWithHooks (/app/node_modules/rea…
GLM 4.7
### The Cause
The error `TypeError: Cannot read properties of undefined (reading 'map')` means that the variable `users` is `undefined` when the component tries to render.
In React, this typically happens for one of two reasons:
1. The **Parent Component** is not passing the `users` prop to `UserList`.
2. The **Parent Component** passed the prop, but the data hasn't loaded yet (e.g., an API call is still pending), so the value is initially `undefined`.
Since `undefined` does not have a `.map()` method, the application crashes.
---
### How to Fix It
There are two main ways to fix this: handling it inside the `UserList` component (defensive coding) or fixing the parent component passing…
42.0s1384 output tokens
GPT-5.6 Terra
The error means `users` is `undefined` when `UserList` renders, so React tries to evaluate:
```js
users.map(...)
```
But `.map()` only exists on arrays.
Common causes:
- The parent component does not pass a `users` prop.
- The user data is loaded asynchronously and has not arrived on the first render.
- The API response shape differs from what you expect, e.g. data is in `response.data.users` rather than `response.data`.
- The parent initializes its user state as `undefined` or `null`.
## Fix options
### 1. Initialize users as an empty array
In the parent component:
```jsx
const [users, setUsers] = useState([]);
```
Then pass it normally:
```jsx
<UserList users={users} />
```
An e…
5.8s559 output tokens
Code Review
Please review the following Python function and suggest improvements for readability, performance, and best practices: def get_data(url, retries=3): import requests import time for i in range(retries):…
GLM 4.7
Here is a review of your function with suggestions for improving readability, performance, and best practices, followed by a refactored version of the code.
### Code Review
1. **Import Placement:** You import `requests` and `time` inside the function. Imports should generally be placed at the top of the file so they are loaded once when the module starts, rather than every time the function is called.
2. **Bare `except:`:** You used `except:` without specifying an exception type. This is dangerous because it catches *everything*, including `KeyboardInterrupt` (Ctrl+C) and `SystemExit`, preventing your program from terminating when needed. You should specifically catch `requests.exception…
24.5s2951 output tokens
GPT-5.6 Terra
Here are the main areas to improve:
### Issues in the original code
- **Imports inside the function**: Move `requests` and `time` imports to module scope.
- **Bare `except:`**: This catches everything, including `KeyboardInterrupt` and programming errors. Catch `requests.RequestException` (and possibly JSON decoding errors) instead.
- **Status code check**: Prefer `response.raise_for_status()` or `response.ok` over manually checking only for `200`. Successful responses can include other 2xx status codes.
- **`== True`**: Use `if item.get("active"):` instead.
- **Potential `KeyError`**: `item["name"]`, `item["id"]`, and `item["active"]` can fail if the API response is incomplete. Use `.get(…
10.6s939 output tokens
Frequently asked questions
- Which is cheaper, GLM 4.7 or GPT-5.6 Terra?
- Per 1M output tokens, GLM 4.7 is ¥393.14 and GPT-5.6 Terra is ¥2,144.4 on FastMetal (yen, before tax), so GLM 4.7 is cheaper.
- How do the context windows of GLM 4.7 and GPT-5.6 Terra compare?
- GLM 4.7 takes 202,752 tokens; GPT-5.6 Terra takes 1,050,000.
- Which ranks higher, GLM 4.7 or GPT-5.6 Terra?
- GPT-5.6 Terra ranks higher on the public arena (GLM 4.7 #79, GPT-5.6 Terra #19). Ranks move as the leaderboard updates.
- Can I use GLM 4.7 and GPT-5.6 Terra with the same API key?
- Yes. On FastMetal's OpenAI-compatible endpoint, switch by passing "glm-4.7" or "gpt-5.6-terra" as the model. Each is billed at its own rate from the same prepaid balance.
Try both on one API key
Create an account and add credit to call GLM 4.7 and GPT-5.6 Terra from the browser chat and the API. No monthly fee.