dispatchAI/sdk-test-results

dispatchAI SDK Test Results

SDK: pip install dispatchai

Test 1: Model Loading

from dispatchai import load_model
model = load_model("SmolLM2-135M-Instruct-mobile", backend="gguf")

Result: ✅ Model loads successfully

Test 2: Chat

response = model.chat("What is the capital of France?")

Result: ✅ "The capital of France is Paris."

Test 3: Recommend

from dispatchai import recommend
rec = recommend(ram_mb=2048, task="chat")

Result: ✅ Returns SmolLM2-135M (smallest, fastest)

Test 4: List Models

from dispatchai import list_models
models = list_models(task="chat")

Result: ✅ Returns 31 working models

Test 5: Cost Calculator

from dispatchai import calculate_cost
result = calculate_cost(daily_queries=10000)

Result: ✅ Returns savings of $18,249/year

Test 6: Latency Estimator

from dispatchai import estimate_latency
lat = estimate_latency("1B", "Q4_K_M")

Result: ✅ Returns 18.2 t/s (matches phone benchmark)

All SDK functions verified working June 2026.

🚀 dispatchAI