See how AI coding assistants really work. Load a repo, enter a prompt, and watch the context window fill up — system prompt, AGENTS.md, /tools, tool calls, and all.
Click a file in the tree to view its contents.
Or try a preset prompt below to watch the agentic loop in action.
Reasoning effort controls how many thinking tokens the model spends before producing its final answer. Higher effort improves accuracy on hard problems at the cost of latency and token spend. Thinking tokens are billed but do not appear in the context bar above — budget for them separately.
| Level | Anthropic effort |
OpenAI reasoning.effort |
|---|---|---|
| Off | — | none — no thinking tokens |
| Low | low — ~1k think tokens |
low |
| Medium | medium |
medium (default for o-series) |
| High | high |
high |
| Max | max — up to 32k think tokens |
xhigh — Extreme Thinking mode |
Anthropic deprecated the explicit budget_tokens parameter in Claude 4.x in favour of this effort enum. The underlying mechanism is the same: the model generates a scratchpad of chain-of-thought tokens that are not returned to the caller but influence the final answer.