CodeSmart
← Compare plans

Claude Max

AnthropicHigh

pro tier · last verified 2026-06-14

Pricing

Monthly
$100/mo
Annual (per month)
—
Billing
Flat
EstimatedSource· 2026-06-14

Max plan: ~5x usage vs Pro. A $200/mo tier (20x Pro) was also advertised.

Models available

Claude Opus 4.8Claude Fable 5Claude Opus 4.7Claude Sonnet 4.6

Observed usage limitsRaw limits as disclosed by the provider. These are the source data — not the estimates below.Read methodology →

As disclosed by the provider, with provenance.

LimitValueApplies toConfidenceSource
messages per period225 messagesAll modelsEstimatedSource· 2026-06-17
unknown—~5x Pro usage at $100/mo tier; ~20x Pro at $200/mo tier. No hard token cap published.all modelsEstimatedSource· 2026-06-14

Estimated valueEstimates, not guarantees: token budgets derived from the observed limits above, then quality-adjusted.Read methodology →

Per-model monthly estimates. Estimated, not guaranteed — every figure carries a confidence level.

Claude Opus 4.8
Intel.—Unknown
Estimated tokens / monthEstimated, not guaranteed. See how token budgets are derived from observed usage limits.Read methodology →
450K tokUnknown
Intelligence-adjusted capacity / month
—Unknown
Value per intelligence per task—
Claude Fable 5
Intel.—Unknown
Estimated tokens / monthEstimated, not guaranteed. See how token budgets are derived from observed usage limits.Read methodology →
450K tokUnknown
Intelligence-adjusted capacity / month
—Unknown
Value per intelligence per task—
Claude Opus 4.7
Intel.—Unknown
Estimated tokens / monthEstimated, not guaranteed. See how token budgets are derived from observed usage limits.Read methodology →
450K tokUnknown
Intelligence-adjusted capacity / month
—Unknown
Value per intelligence per task—
Claude Sonnet 4.6
Intel.—Unknown
Estimated tokens / monthEstimated, not guaranteed. See how token budgets are derived from observed usage limits.Read methodology →
450K tokUnknown
Intelligence-adjusted capacity / month
—Unknown
Value per intelligence per task—
How Value per Intelligence per Task is calculatedThe full formula and constants are documented in the methodology.Read methodology →Show steps
  1. 1
    Intelligence Score—

    50% AA agentic + 40% AA coding + 10% AA speed.

  2. 2
    Estimated monthly tokens450K

    Normalized from the plan's observed usage limits — an estimate, not a guarantee.

  3. 3
    Model-adjusted tokens—

    Tokens scaled for the model's relative efficiency.

  4. 4
    Intelligence-adjusted capacity—

    Tokens × (Intelligence Score / 100).

  5. 5
    Per dollar$100/mo

    Intelligence-adjusted capacity ÷ monthly price = raw value.

  6. 6
    Raw value—

    Intelligence-adjusted capacity per dollar before normalization.

  7. 7
    Normalized Value per Intelligence per Task—

    Scaled to 0–100 against a reference of 1M quality-adjusted tokens at $20/mo.

Caveats
  • •No AA data available
  • •No cost-per-task data — efficiency neutral