Instructions to use dcostenco/prism-coder-32b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dcostenco/prism-coder-32b with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="dcostenco/prism-coder-32b",
	filename="prism-coder-32b-q4km.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use dcostenco/prism-coder-32b with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf dcostenco/prism-coder-32b
# Run inference directly in the terminal:
llama-cli -hf dcostenco/prism-coder-32b

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf dcostenco/prism-coder-32b
# Run inference directly in the terminal:
llama-cli -hf dcostenco/prism-coder-32b

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf dcostenco/prism-coder-32b
# Run inference directly in the terminal:
./llama-cli -hf dcostenco/prism-coder-32b

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf dcostenco/prism-coder-32b
# Run inference directly in the terminal:
./build/bin/llama-cli -hf dcostenco/prism-coder-32b

Use Docker

docker model run hf.co/dcostenco/prism-coder-32b

LM Studio
Jan

vLLM

How to use dcostenco/prism-coder-32b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dcostenco/prism-coder-32b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dcostenco/prism-coder-32b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/dcostenco/prism-coder-32b

Ollama
How to use dcostenco/prism-coder-32b with Ollama:
```
ollama run hf.co/dcostenco/prism-coder-32b
```

Unsloth Studio new

How to use dcostenco/prism-coder-32b with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for dcostenco/prism-coder-32b to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for dcostenco/prism-coder-32b to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for dcostenco/prism-coder-32b to start chatting

Pi new

How to use dcostenco/prism-coder-32b with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf dcostenco/prism-coder-32b

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "dcostenco/prism-coder-32b"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use dcostenco/prism-coder-32b with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf dcostenco/prism-coder-32b

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default dcostenco/prism-coder-32b

Run Hermes

hermes

Docker Model Runner
How to use dcostenco/prism-coder-32b with Docker Model Runner:
```
docker model run hf.co/dcostenco/prism-coder-32b
```

Lemonade

How to use dcostenco/prism-coder-32b with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull dcostenco/prism-coder-32b

Run and chat with the model

lemonade run user.prism-coder-32b-{{QUANT_TAG}}

List all available models

lemonade list

dcostenco commited on 15 days ago

Commit

6b0fc59

verified ·

1 Parent(s): 0266b49

Update model card with training details, cascade position, and file table

Browse files

Files changed (1) hide show

README.md +28 -74

README.md CHANGED Viewed

@@ -1,96 +1,50 @@
 ---
-base_model: Qwen/QwQ-32B
-library_name: peft
-pipeline_tag: text-generation
-license: apache-2.0
 language: en
 tags:
-- lora
-- sft
-- function-calling
-- tool-use
-- mcp
-- aac
-- prism-coder
 ---
-# prism-coder:32b (v19) — 97.3% routing accuracy
-LoRA fine-tune of **Qwen/QwQ-32B** for offline MCP tool routing.
-## Routing accuracy — 100-case Prism eval (May 15 2026, 3-seed mean)
-| Category | Score |
-|---|---|
-| **Overall** | **97.3% ± 0.6%** |
-| All 7 MCP tools | 100% each |
-| AAC plain-text | ~90% |
-| translate | 83% |
-| edge (multi-step) | 100% |
-| avg latency | 2.4s |
-| invented tools | 0 |
-Uses `nothink` template + v27 system prompt with labeled category headers.
-## Usage
-```bash
-ollama pull dcostenco/prism-coder:32b
-```
-## Hardware
-- **Mac**: M2 Ultra+ / 48GB+
-- **Linux**: A100 40GB+
-- **VRAM**: ~22 GB
----
-### All Prism Coder models
-| Model | Accuracy | Size | Device | HuggingFace |
-|---|---|---|---|---|
-| **prism-coder:14b** | **98%** | 8.4 GB | Mac / iPad Pro 16GB | [dcostenco/prism-coder-14b](https://huggingface.co/dcostenco/prism-coder-14b) |
-| **prism-coder:8b** | **96%** | 4.7 GB | iPhone / iPad 8GB | [dcostenco/prism-coder-8b](https://huggingface.co/dcostenco/prism-coder-8b) |
-| **prism-coder:32b** | **97.3%** | 19 GB | Mac M2 Ultra+ | [dcostenco/prism-coder-32b](https://huggingface.co/dcostenco/prism-coder-32b) |
-| **prism-coder:1.7b** | **88%** | 2.2 GB | Any device / iPhone | [dcostenco/prism-coder-1.7b](https://huggingface.co/dcostenco/prism-coder-1.7b) |
-GitHub: [dcostenco/prism-coder](https://github.com/dcostenco/prism-coder) · AAC app: [dcostenco/prism-aac](https://github.com/dcostenco/prism-aac) · Portal: [synalux.ai](https://synalux.ai)
-## Get the full stack
-The model routes tool calls — but needs a backend to route TO:
 ```bash
-# Install the memory server (free, local, no API keys)
-npm install -g prism-mcp-server
-# Pull the model
-ollama pull dcostenco/prism-coder:32b
-# Done — your AI agent now has persistent memory + 98% tool routing
 ```
-**Free tier:** local SQLite, no cloud, no account needed.
-**Synalux portal:** cloud sync, HIPAA dashboard, team access, Claude fallback → [synalux.ai](https://synalux.ai)
----
-## Prism Routing Benchmark
-This model is evaluated on the [Prism Routing Benchmark](https://github.com/dcostenco/prism-coder/tree/main/tests/benchmarks/prism-routing-100) — a 100-case, 13-category eval for MCP tool routing. Run it yourself:
-```bash
-git clone https://github.com/dcostenco/prism-coder
-cd prism-coder
-python3 tests/benchmarks/prism-routing-100/benchmark.py --models 32b --seed 2027
-```
-Not a general function-calling benchmark (BFCL). This measures routing precision on 7 specific MCP tools — the task these models were built for. The value is **offline reliability at zero cost**, not competing with frontier models on arbitrary APIs.
-## License
-Apache-2.0.

 ---
 language: en
+license: apache-2.0
+base_model: Qwen/Qwen3-32B
 tags:
+  - tool-calling
+  - routing
+  - coding
+  - aac
 ---
+# prism-coder:32b — AAC Tool Router + Coder (32B)
+Fine-tuned from **Qwen3-32B** for tool routing and advanced code assistance in the [Prism AAC](https://github.com/dcostenco/prism-aac) system.
+**BFCL accuracy: 99%** on 100-case routing benchmark. Quality escalation tier in the desktop cascade — catches the ~1-3% of cases where 14B is uncertain.
+## What it does
+- Perfect tool routing on all tested categories
+- Advanced code generation and architecture assistance
+- Complex multi-step session management
+- Final local quality gate before cloud Claude
+## Deployment
+Available on **Ollama Hub** (recommended — avoids 18GB download for Ollama users):
 ```bash
+ollama run dcostenco/prism-coder:32b
 ```
+Or pull manually with the GGUF from this repo when available.
+## Cascade position
+Desktop cascade: **14B → 32B (escalation) → cloud Claude**
+When 14B returns low-confidence or fails, 32B is invoked automatically. Users with Ollama running get 32B as their local ceiling before cloud.
+## Training
+- **Base**: Qwen3-32B
+- **Method**: MLX LoRA fine-tuning (v28-codebase + routing)
+- **Hardware**: Apple Silicon (M-series, 64GB RAM)
+- **Eval**: BFCL routing 99% (11/11 on manual benchmark)
+## Note on GGUF
+The full Q4_K_M GGUF is 18GB. It is distributed via **Ollama Hub** at `dcostenco/prism-coder:32b` to avoid large download overhead. Direct GGUF will be added here in a future release.