Free course · Registered members · Lesson 5 / 5
Choose with a test instead of chasing every release
Finish with a repeatable comparison procedure for changes in models and pricing.
Editorial team: Take Two · Published
Lesson preview
Prepare three cases
Use a short rewrite, extraction from incomplete notes and research with official sources. Keep data and criteria identical. Research requires web access; record its absence rather than comparing an offline model with a browsing product as equivalent.
Use a simple rubric
Score 0 = unmet, 1 = met after corrections, 2 = met without substantial corrections. Evaluate five criteria: factual accuracy, format, handling missing facts, verifiable sources and review time under your chosen threshold. Maximum score is 10, but a critical error can still disqualify a result.
What to check as you practise
- Did you use identical prompts and data?
- Did you record corrections and limits, not only totals?
- Does the decision state when to reconsider price and features?
Continue with the full lesson
Register free to access the complete explanation, examples, exercise and commented answer. This preview stays open to everyone.
Sources and further reading
Official sources checked on 6 October 2026. This is a dated selection, not a full catalogue or an automatically updated price list. Follow current vendor terms and local checkout prices.
- ChatGPT — piani ↗
- ChatGPT Plus — prezzo e API ↗
- ChatGPT Pro — fasce di abbonamento ↗
- ChatGPT Go — disponibilità ↗
- OpenAI — modelli in ChatGPT ↗
- ChatGPT Business — piani e posti ↗
- Claude — piani ↗
- Claude Max — prezzo ↗
- Anthropic — modelli API ↗
- Google — piani AI ↗
- Google — modelli Gemini API ↗
- Microsoft — Copilot per privati ↗
- Perplexity — prezzi pubblici ↗
- Perplexity — piani ↗
- Perplexity Max — prezzo ↗
- Mistral — piani ↗
- Mistral — da Le Chat a Vibe ↗
- Mistral — modelli ↗