Maximize value from your Vigthoria subscription with smart usage strategies.
Monitor your usage patterns to identify optimization opportunities:
Export your usage data monthly to identify trends. Many users find 20-30% of their usage is redundant or could use lighter models.
Not every task needs the most powerful model:
| Task Type | Recommended Model | Cost Level |
|---|---|---|
| Simple Q&A, classification | vigthoria-reasoning-v2 | Standard |
| Code generation | vigthoria-code-v2 | Standard |
| Creative content | vigthoria-creative-v2 | Standard |
| Image analysis | vigthoria-vision-v2 | Premium |
By matching models to tasks instead of using one model for everything.
Shorter, clearer prompts use fewer input tokens:
// Before: 45 tokens "I would really appreciate it if you could please help me by writing a function that takes a number as input and returns whether that number is a prime number or not." // After: 18 tokens "Write a function isPrime(n) that returns true if n is prime."
Set appropriate max_tokens for each use case:
{
"max_tokens": 200, // For short answers
"max_tokens": 500, // For explanations
"max_tokens": 1500 // For articles
}
End generation early when you have what you need:
{
"stop": ["---", "END", "\n\n\n"]
}
By reducing average tokens per request from 2000 to 800.
Don't pay for the same generation twice:
import hashlib
import redis
cache = redis.Redis()
def cached_generation(prompt, model, **kwargs):
# Create cache key from request
cache_key = hashlib.sha256(
f"{model}:{prompt}:{kwargs}".encode()
).hexdigest()
# Check cache
cached = cache.get(cache_key)
if cached:
return json.loads(cached)
# Generate and cache
response = vigthoria.chat.completions.create(
model=model,
messages=[{"role": "user", "content": prompt}],
**kwargs
)
# Cache for 24 hours
cache.setex(cache_key, 86400, json.dumps(response))
return response
Good candidates for caching:
Depending on how many repeated requests you have.
For non-real-time tasks, batch requests during off-peak hours:
// Instead of 10 separate requests:
const items = ['item1', 'item2', 'item3', ...];
// Combine into one:
const response = await vigthoria.chat.completions.create({
model: 'vigthoria-reasoning-v2',
messages: [{
role: 'user',
content: `Analyze these 10 items and provide a summary for each:
${items.join('\n')}`
}]
});
Prevent surprise overages with proactive monitoring:
Review your plan quarterly:
Annual plans typically offer 15-20% savings over monthly billing. If you're committed to Vigthoria, consider switching.