all five, side by side

Four hosted assistants, plus Qwen for running it on your own device. Tap a card to circle the one you’re eyeing.

prices from a Sept 28, 2026 snapshot. Double-check before you buy!
API prices are per 1M tokens, input / output
who it’s for (Gemini) →
who it’s for (Claude) →
who it’s for (ChatGPT) →
who it’s for (DeepSeek) →
who it’s for (Qwen) →
remember: cost per token isn’t cost per finished task. A cheap model that retries eight times can cost more than a pricier one that gets it right once.This is my opinion from one research snapshot, not a lab benchmark.
still unsure? take the quiz →