Best fit
Complex implementation and review loops where code reasoning beats raw speed.
Give Claude Code the change that needs careful judgment.
Claude Code takes the subtle implementation, the review-heavy change and the follow-up that turns on a tradeoff. ProPR plans it, runs it in isolation and returns it as a pull request with the full log attached.
Approved planContext and scope attached
Opus / Fable / Sonnet
Isolated worker
GitHub pull requestCommits, logs, and follow-ups
Your credentials
Task history kept
GitHub merge decision
Agent fit and access
Claude Code plans and models
Claude Pro at $20 a month includes Claude Code with Opus, Sonnet and Haiku under a five-hour session limit and a weekly limit. Max at $100 multiplies that usage by five and adds Fable; Ultra at $200 gives ten times the Pro weekly usage with no five-hour limit on verified accounts. Anthropic publishes the multipliers and no message or token counts.
| Plan | Price (USD) | Models | Published limits |
|---|---|---|---|
| Claude Free | $0 | Sonnet, Haiku. | Claude Code is excluded. |
| Claude Pro | $20 / month ($17 billed annually) | Opus, Sonnet, Haiku; Fable through usage credits. | Five-hour session limit and weekly limit. No number published. |
| Claude Max | $100 / month | Opus, Sonnet, Haiku; Fable up to 50% of the weekly limit. | 5× Pro usage, five-hour and weekly. |
| Claude Ultra | $200 / month | Opus, Sonnet, Haiku; Fable up to 50% of the weekly limit. | 10× Pro weekly usage; no five-hour limit on verified accounts. |
| Claude Team | $25 standard / $125 premium per seat / month ($20 / $100 billed annually) | Opus, Sonnet, Haiku; Fable on premium seats. | More than Pro; premium seats get 5× a standard seat. |
| Anthropic API key | Pay-as-you-go per token | Every Claude API model. | Metered billing with no plan cap. |
Every subscription runs headless: claude -p runs one task on the subscription login and exits, which is how ProPR drives it.
Sources: Claude pricing, Claude Code on Pro and Max, Claude Code headless docs. Verified November 2026.
That subscription runs in ProPR on GitHub issues and pull requests. Label an issue llm-claude-opus55 or llm-claude-fable51 and ProPR plans the task, runs Claude Code on your Claude login in an isolated workspace, and opens the PR. You pay Anthropic directly; ProPR adds no token markup. The guide to running Claude Code on GitHub issues and pull requests walks through the first task.
Access
A Claude subscription or an Anthropic API key.
Role in ProPR
Claude plans, implements, reviews and handles PR fine-tuning, all through the same review path.
Pick the Claude that fits the phase.
Configure the models once and use them everywhere: issue labels, plan settings, /switch and /review. Set the reasoning level globally, per model, or per issue with a level-* label, from low to max, including Claude Code's adaptive auto.
| Model | Issue label | Context | Notes |
|---|---|---|---|
| Claude Opus 5.5 (default) | llm-claude-opus55 | 1M | ProPR's default Claude model for planning, implementation, and review. |
| Claude Fable 5.1 / Fable 5 | llm-claude-fable51 / -fable5 | 1M | The hardest planning, architecture, and review work. |
| Claude Sonnet 5.5 | llm-claude-sonnet55 | 1M | Routine implementation and PR follow-ups at a faster pace. |
| Claude Haiku 5.5 | llm-claude-haiku55 | 1M | Fast, high-volume passes: summaries, triage, and quick checks. |
| Claude Opus 5 / Sonnet 5 | llm-claude-opus5 / -sonnet5 | 1M | The previous 5-series generation, pinned for workflows tuned against it. |
| Claude 4.x series | llm-claude-opus48, -sonnet46, -haiku45, … | 200K–1M | Opus 4.8 through Haiku 4.5, under the earlier-models fold in the picker. |
Model choice
Pick the Claude model for the job and give it a readable alias for the team.Phase routing
Route Claude to planning and review, the phases that reward careful reasoning, and mix in faster models for routine implementation.Native goals
Claude Code runs Goals in its own goal mode, and your corrections reach the running session live.Usage tracking
With Agent Tank, ProPR shows the Claude capacity each call spent.
Set up Claude Code
Claude Code setup, models and credentials in the ProPR docs → ProPR runs the official Claude Code CLI (@anthropic-ai/claude-code). Using Claude Code on its own today? Read the direct Claude Code comparison.