Skip to content

chore(pricing): add Fable 5, align Opus rates to current 5/25 USD - #28

Open
jlwainwright wants to merge 1 commit into
nateherkai:mainfrom
jlwainwright:chore/pricing-fable5-opus-update
Open

chore(pricing): add Fable 5, align Opus rates to current 5/25 USD#28
jlwainwright wants to merge 1 commit into
nateherkai:mainfrom
jlwainwright:chore/pricing-fable5-opus-update

Conversation

@jlwainwright

Copy link
Copy Markdown

What

Updates pricing.json with current Anthropic rates.

Model Input Output Cache read Cache write 5m / 1h
Fable 5 (new) $10 $50 $1 $12.50 / $20
Opus 4.8 (new entry) $5 $25 $0.50 $6.25 / $10
Opus 4.7 $5 $25 (was $15/$75) $0.50 $6.25 / $10
Opus 4.6 $5 $25 (was $15/$75) $0.50 $6.25 / $10

Sonnet 4.6 and Haiku 4.5 unchanged.

Why

  • claude-fable-5 is referenced in the model list (and is the new top tier) but had no pricing entry → costs showed $0 for Fable traffic.
  • Opus 4.6/4.7 were priced at the older $15/$75; current list is $5/$25.

Source

Rates taken from anthropic.com/pricing.

Caveats

  • 5m/1h cache split: the pricing page lists a single "Prompt caching Write" rate per model. I used the standard 1.25×/2× (5m/1h) ratio that the existing entries follow. If Anthropic publishes distinct 5m/1h write rates, happy to adjust.
  • Opus 4.6/4.7 aren't separately listed on the pricing page; I extended the 4.8 rate. If older Opus generations had different pricing, let me know.

Related: #14 (closed) covered the Opus rate change but predated Fable 5.

Fable 5: input $10, output $50 per M tokens.
Cache: read $1, write $12.50 (5m) / $20 (1h) per M tokens.

Opus 4.6/4.7/4.8: $5 input / $25 output (was $15/$75 in repo).
Matches rates currently listed at anthropic.com/pricing.

Source: https://www.anthropic.com/pricing
muckybuzzwoo added a commit to muckybuzzwoo/token-dashboard that referenced this pull request Jul 7, 2026
…as deliberately skipped

PR nateherkai#28 (Fable 5 + Opus 5/25 pricing) is fully superseded by existing entries incl. Sonnet 5. PR nateherkai#27 (configurable scan interval) is server-only and would desync the fork's coupled server/client 60s scan cadence.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant