Chat model
Qwen3.5 4B on Apple silicon Macs
In testing, opens soon
Quick everyday chat and short answers. Smarty Pants serves a pinned, quantized 4-bit MLX build on qualified Macs with 16 GB of memory or more, through an OpenAI-compatible API.
$0.10
$0.10 per million input tokens, $0.25 per million output tokens
Listed price on November 19, 2026. Current price: GET /v1/models.
Qwen3.5 4B at a glance
What Smarty Pants serves, from its public catalog, and what the model's developer publishes about it.
- API name
smartypants/qwen3.5-4b- Endpoint
- POST /v1/chat/completions
- Build served
- 4-bit MLX (qwen3.5-4b-mlx-4bit-r32f3e8e)
- Mac memory needed
- 16 GB or more
- Context on Smarty Pants
- 8,192 tokens
- Max output
- 2,048 tokens
- Developer
- Qwen (Alibaba Cloud)
- Parameters
- 4B
- License
- Apache-2.0
- Upstream context
- 262,144 tokens (extensible up to 1,010,000)
- First released
- February 2026
- Status
- In testing, opens soon
Smarty Pants catalog checked November 19, 2026. Upstream facts from the developer's model card (sources below).
What Qwen3.5 4B is good for
Multimodal (text, image, video input) chat model for reasoning, coding, agent and visual-understanding tasks, with coverage of 201 languages and dialects. Runs in thinking mode by default; thinking can be turned off for direct answers.
Same 32-layer hybrid layout as Qwen3.5-9B but with a 2,560-wide hidden dimension and an output layer tied to the token embedding, which keeps it at 4B parameters while supporting 201 languages and dialects.
The upstream model also accepts images. Smarty Pants serves it for text in, text out.
Use it for
- Short answers and classification
- High-volume, simple text tasks
Look elsewhere for
- Inputs longer than 8,192 tokens
- Multi-step reasoning and harder code
Measured on a real Mac
The standard 4-bit build Smarty Pants serves ran at 99.7 tokens a second with 4.13 GB peak memory on an Apple M5 Pro Mac mini with 64 GB, and answered 89.0% of 500 GSM8K math problems correctly with thinking off.
Measured 2026-11-05 by the Smarty Pants model factory. Full method and results: Smarty Pants Mac benchmarks on Hugging Face. Live speed depends on the Mac serving your request.
How to call Qwen3.5 4B
Smarty Pants works with OpenAI's libraries: change the base URL and the key, and keep the rest. Get a key at smartypants.network/api.
from openai import OpenAI
client = OpenAI(base_url="https://smartypants.network/v1", api_key="YOUR_SMARTY_PANTS_KEY")
reply = client.chat.completions.create(
model="smartypants/qwen3.5-4b",
messages=[{"role": "user", "content": "Hello!"}],
)
print(reply.choices[0].message.content) Before sending real traffic, check is_ready for this model on GET /v1/models. A 429 or 503 means no Mac can take the request right now: wait the Retry-After seconds or use your fallback provider.
Which Macs can serve Qwen3.5 4B
Smarty Pants serves this model only on Macs with 16 GB of unified memory or more. These Apple silicon chips were sold with at least one memory size that big (memory options from Apple's tech specs).
| Chip | Memory sizes that fit | Memory sizes too small |
|---|---|---|
| M1 | 16 GB | 8 GB |
| M1 Pro | 16 GB, 32 GB | None |
| M1 Max | 32 GB, 64 GB | None |
| M1 Ultra | 64 GB, 128 GB | None |
| M2 | 16 GB, 24 GB | 8 GB |
| M2 Pro | 16 GB, 32 GB | None |
| M2 Max | 32 GB, 64 GB, 96 GB | None |
| M2 Ultra | 64 GB, 128 GB, 192 GB | None |
| M3 | 16 GB, 24 GB | 8 GB |
| M3 Pro | 18 GB, 36 GB | None |
| M3 Max | 36 GB, 48 GB, 64 GB, 96 GB, 128 GB | None |
| M3 Ultra | 96 GB, 256 GB, 512 GB | None |
| M4 | 16 GB, 24 GB, 32 GB | None |
| M4 Pro | 24 GB, 48 GB, 64 GB | None |
| M4 Max | 36 GB, 48 GB, 64 GB, 128 GB | None |
| M5 | 16 GB, 24 GB, 32 GB | None |
| M5 Pro | 24 GB, 48 GB, 64 GB | None |
| M5 Max | 36 GB, 48 GB, 64 GB, 128 GB | None |
| M5 Ultra | 96 GB, 256 GB, 512 GB | None |
| M6 | 16 GB, 24 GB, 32 GB | None |
Frequently asked questions
What is the API name for Qwen3.5 4B on Smarty Pants?
Use "smartypants/qwen3.5-4b" as the model with the OpenAI-compatible base URL https://smartypants.network/v1. The older name smartypants/smarty still works.
How much does Qwen3.5 4B cost on Smarty Pants?
The listed price on November 19, 2026 was $0.10 per million input tokens, $0.25 per million output tokens. Prices can change; GET https://smartypants.network/v1/models returns the current price, and you are not charged for a request Smarty Pants can't serve.
How much memory does a Mac need to serve Qwen3.5 4B?
Smarty Pants serves this 4-bit MLX build only on qualified Apple silicon Macs with 16 GB of unified memory or more.
What context length does Qwen3.5 4B have on Smarty Pants?
Smarty Pants accepts up to 8,192 tokens per request and writes up to 2,048. The upstream model card lists 262,144 tokens (extensible up to 1,010,000); Smarty Pants serves a smaller window so the model fits on Macs.
Is Qwen3.5 4B available right now?
On November 19, 2026 it was in testing and opens soon. How many Macs have it online changes through the day; check is_ready on https://smartypants.network/v1/models before sending real traffic, and keep a fallback provider.
Can I send private data to Qwen3.5 4B on Smarty Pants?
Requests on the general network run on Macs other people own, and a serving Mac decrypts a request in memory to answer it. Keep secrets, personal and confidential data off the general network; companies can run work on their own Macs with a Group.
Sources
Other chat models
Try a pinned open model on real Macs.
See every model, its listed price and the code to call it. Smarty Pants works with the OpenAI SDK you already use: change the base URL and the key.
Requests run on Macs owned by other people. Owners agree not to read requests and the app doesn't show them, but a serving Mac decrypts each request in memory to answer it, so that is a rule and a design, not a technical guarantee. Don't send sensitive data on the general network.