GLM 5.3 vs DeepSeek V4 Pro
GLM 5.3 and DeepSeek V4 Pro are both large models with published weights, used for heavy text work. The 0813 build is the general-availability release of DeepSeek V4 Pro.
Today's prices and specs
Prices are in Rial per million tokens and specs are read straight from the IranRouter catalogue. The best value in each row is green. The intelligence and coding indexes come from public benchmarks, and hovering over a speed or first-token figure shows where it was measured.
| Z.ai: GLM 5.3Z.ai | DeepSeek: DeepSeek V4 Pro 0813DeepSeek | |
|---|---|---|
| Input (IRR / 1M tokens) | 191,947 | 3,624,747 |
| Output (IRR / 1M tokens) | 3,844,429 | 10,874,241 |
| One message (IRR) | 2,306 | 12,687 |
| Context (tokens) | 1M | 1M |
| Max output (tokens) | 944K | 944K |
| Intelligence | 31.7 | 23.1 |
| Coding | — | — |
| Speed (tokens/s) | 86From public market statistics for this model | 47From public market statistics for this model |
| First token (s) | 0.8From public market statistics for this model | 1.0From public market statistics for this model |
| Health, 30 days | 91.70% | 50.00% |
| Usage rank | #10 | #39 |
| Can do | Tool callingJSON outputReasoning | Tool callingJSON outputReasoning |
| Released | Aug 19, 2026 | Aug 12, 2026 |
The differences at a glance
- DeepSeek V4 Pro 0813's input price is about 18.9 times GLM 5.3's.
- DeepSeek V4 Pro 0813's output price is about 2.8 times GLM 5.3's.
- One typical message costs about 2,306 IRR on GLM 5.3 and about 13K IRR on DeepSeek V4 Pro 0813.
- Both hold about 1M tokens of context.
- Both accept text as input.
When to choose GLM 5.3
- Complex software engineering and long agent tasks, which its maker names.
- Work that has to keep a lot of text in view at once.
When to choose DeepSeek V4 Pro
- When your prompts are already tuned for DeepSeek models.
- Text work and general reasoning that need no images or files.
Before you decide
The balance of input and output in your work moves the cost a lot, because each model sets those two rates separately. Read the table against a real sample of your usage.
from openai import OpenAI client = OpenAI(api_key="ir-...", base_url="https://iranrouter.com/v1") for model in ["z-ai/glm-5.3", "deepseek/deepseek-v4-pro-0813"]: r = client.chat.completions.create(model=model, messages=[{"role": "user", "content": "..."}]) print(model, r.choices[0].message.content)
Frequently asked questions
Which is cheaper, GLM 5.3 or DeepSeek V4 Pro?
At today's rates GLM 5.3 is cheaper on both input and output. Rates move with the daily price, so the table on this page is the reference.
How does DeepSeek V4 Pro 0813 differ from the earlier V4 Pro?
0813 is the general-availability release, and the two are listed separately. Each one's price and scores are on its own page.
Can I try both on one account?
Yes. One ir- key and the same Rial wallet reach both, and switching means changing only the model value in the request. You can set a spending limit on each key.

