Install Ollama on Mac
Install the official Mac app, open it, then start Terminal.
Official Ollama downloadAgents / WORKFLOW
Build a local AI tool agent with Qwen3 8B
Run a bounded local function through Qwen3 8B and inspect the tool call, tool result, and final answer.
EstimatedSource checked 9/29/2026
LOCALRENTED GPU
Before you begin
Hardware
Estimated memory starting points for the 5.2 GB Qwen3 8B Ollama package with one tool and 4K context. Runtime and KV cache need extra memory beyond the download.
| Machine | Qwen3 · 8B |
|---|---|
| NVIDIA GPU · 8 GB | Possible with CPU offload and 16 GB-plus system RAM; check ollama ps. |
| NVIDIA GPU · 12 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| NVIDIA GPU · 16 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| NVIDIA GPU · 24 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| NVIDIA GPU · 32 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| AMD Radeon · 8 GB | Possible with CPU offload and 16 GB-plus system RAM; check ollama ps. |
| AMD Radeon · 16 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| AMD Radeon · 24 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| AMD Radeon · 32 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| Intel Arc · 8 GB | Possible with CPU offload and 16 GB-plus system RAM; check ollama ps. |
| Intel Arc · 16 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| Intel Arc · 24 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| Intel Arc · 32 GB | Runs at 4K if Ollama loads the package on this GPU; check ollama ps. |
| Apple Silicon · 16 GB unified | Runs at 4K with room for one tool; longer history uses more memory. |
| Apple Silicon · 24 GB unified | Runs at 4K with room for one tool; longer history uses more memory. |
| Apple Silicon · 32 GB unified | Runs at 4K with room for one tool; longer history uses more memory. |
| Apple Silicon · 64 GB unified | Runs at 4K with room for one tool; longer history uses more memory. |
| CPU-only · 16 GB RAM | Can run from system RAM at 4K; expect slower tool turns than GPU. |
| CPU-only · 32 GB RAM | Can run from system RAM at 4K; expect slower tool turns than GPU. |
| CPU-only · 64 GB RAM | Can run from system RAM at 4K; expect slower tool turns than GPU. |
These are memory-fit starting points, not performance tests. Check the exact AMD or Intel Ollama backend and use ollama ps to see the GPU/CPU split; longer context or image input needs more memory.
LOCAL AGENT / OLLAMASOURCE-LINKED SETUP
Let Qwen3 request one safe local tool.
Install the Qwen3 8B Ollama package, run a small Python tool loop, and check the returned value against the local lookup. Use the Windows, Mac, or Linux route for the same task.
Choose a package Run its commands Check the result
01
BEFORE YOU BEGIN
The documented package
This setup uses the package documented for this task. Review its source and supported platforms before starting.
Save your machine in My Hardware to get an automatic starting choice. You can always choose any package yourself.
CURRENT SETUP
View weight source Qwen3 · 8B · official
The Qwen3 8B package downloads 5.2 GB. Start with 4K context. An 8 GB GPU may need CPU offload; 12 GB GPU or 16 GB Apple Silicon unified memory is an estimated smoother starting point. Check ollama ps rather than assuming speed from weight size.
02
THE WORKFLOW
Set up Qwen3 · 8B on your machine
Choose where you want to run it. The instructions below update to match your selection.
CHOOSE YOUR SETUPMac
Use Terminal on Apple Silicon. Ollama for macOS is available from the official download page.
Check the runtime and Python
In Terminal, confirm Ollama and Python 3 are available. Install Python 3 first if python3 is not found.
Install Python 3ollama --version
python3 --versionDownload qwen3:8b
Pull the exact package before running the example, so model download failures are separate from Python errors.
ollama pull qwen3:8bInstall the local Python libraries
The official Ollama Python client connects to the service on your own machine.
python3 -m pip install --upgrade ollamaCreate tool_agent.py
In Terminal, paste this block. It writes the complete small example as tool_agent.py in your current folder. Read the code before running it.
cat > tool_agent.py <<'PY'
from ollama import chat
def get_temperature(city: str) -> str:
"""Return the temperature from this local example lookup, not live weather."""
return {"New York": "22°C", "London": "15°C"}.get(city, "No sample value")
messages = [{"role": "user", "content": "Using your tool, what is the sample temperature in New York?"}]
response = chat(model="qwen3:8b", messages=messages, tools=[get_temperature], think=True)
messages.append(response.message)
if not response.message.tool_calls:
raise RuntimeError("No tool call: check that qwen3:8b loaded and retry the exact question")
call = response.message.tool_calls[0]
if call.function.name != "get_temperature" or call.function.arguments.get("city") != "New York":
raise ValueError("Unapproved tool name or city; no function was executed")
result = get_temperature(**call.function.arguments)
print("Local tool result:", result)
messages.append({"role": "tool", "tool_name": "get_temperature", "content": result})
final = chat(model="qwen3:8b", messages=messages, tools=[get_temperature], think=True)
print("Final answer:", final.message.content)
PYExecute the tool loop
Run the Python script. It checks that the requested function is exactly get_temperature and the city is New York, executes the local lookup, returns a tool message, then asks Qwen3 for its final answer.
python3 tool_agent.pyCheck the local result and placement
Look for Local tool result: 22°C and a final answer that uses 22°C. If the model does not request the tool, the script raises an error rather than pretending it succeeded. Use ollama ps to see CPU/GPU placement.
ollama ps03
SUCCESS CHECK · Qwen3 · 8B
A completed local tool round trip
The Python output prints Local tool result: 22°C, then Qwen3 uses that tool result in its final answer. The function is only called after its name and arguments pass the allowlist. No shell or network tool is available.
The lookup is a deliberately static sample, not live weather. Never allow an AI-requested tool name or argument to execute before checking it. Start with 4K context; larger tool histories increase memory use.
When it doesn’t go to plan
No tool call returned.
Confirm qwen3:8b is installed, then rerun the exact sample question with tools=[get_temperature]. The script stops instead of inventing a tool result.
Unapproved function or arguments.
Keep the allowlist check. Do not execute unexpected tools; inspect the call and retry with a narrower prompt.
The second response ignores 22°C.
Confirm the assistant tool-call message and role="tool" result are both appended before the second chat request; compare the printed local value with the final answer.
REFERENCE LIBRARY
4 SOURCESSources and files
Original instructions, model files, and compatibility notes behind this setup.
The model behind this workflow
Qwen3 8BKeep exploring
Text
SETUP GUIDEQwen3.5 · 9B private chat setup
Source-backed package and platform-route plan for private local chat with Qwen3.5 · 9B. Preserves the official Ollama 9B package, a smaller 4B family fallback, and a clearly labeled community uncensored GGUF option.
Freshness 38 out of 100. Release age: 254 days old, Feb 27, 2026. Measures age, not quality.
Popularity 91 out of 100. 6,321,700 downloads in 30 days; 2,329 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.
SOFTWAREOllama
Coding
SETUP GUIDEBuild a private coding assistant with Qwen3.8 27B
Run Qwen3.8 27B with Ollama, connect it to Cline in VS Code, and make one reviewable code change in a tiny local project.
Freshness 69 out of 100. Repository age: 95 days old, Aug 5, 2026. Measures age, not quality.
Popularity 96 out of 100. 7,813,215 downloads in 30 days; 19,312 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.
Arena Open Source Web coding: rank 12 of 62. Ranked model: Qwen3.8-27B. Original model ranking, not a benchmark of this package. Export dated Nov 6, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.+3
SOFTWAREOllama + Visual Studio Code + Cline + Python
Images
SETUP GUIDECreate images locally with FLUX.2 Klein
Generate images from text prompts in ComfyUI with the fast distilled or more adjustable base 4B model. Save your results and reuse the visual workflow.
Freshness 32 out of 100. Release age: 297 days old, Jan 15, 2026. Measures age, not quality.
Popularity 77 out of 100. 435,063 downloads in 30 days; 1,097 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.
SOFTWAREComfyUI