On September 1, 2026, Anthropic listed Claude Fable 5.1 as its most capable generally available model for coding and knowledge work. The API price on the Anthropic pricing page is $10 per million input tokens and $50 per million output tokens. Cache reads are $0.25 per million. Cache writes are $12.50 per million. Those figures are USD as published. CAD is on the invoice. This desk is not inventing a Canadian list price.
What Anthropic listed on September 1
The Fable product page dated September 1, 2026, is the primary note. Anthropic’s newsroom listing for that day is titled “Introducing Claude Fable 5.1 and Claude Mythos 5.1.” This briefing does not attach an invented newsroom slug. Fable 5.1 is available to Pro, Max, Team, and Enterprise users. Anthropic also lists it on the Claude Platform and through Amazon Web Services, Google Cloud, and Microsoft Foundry.
“Our most capable model for coding and knowledge work, with research capabilities that offer an early glimpse of how AI will soon contribute to scientific progress.”Anthropic, Introducing Claude Fable 5.1, September 1, 2026
The models overview gives the Claude API id claude-fable-5-1, a 1M-token context window, and 128K max output. Anthropic’s lineup prices on that page match the API cards: Opus 5 at $5 / $25, Sonnet 5 at $2 / $10, and Haiku 4.5 at $1 / $5 per million input and output tokens.
What it costs
These are Anthropic’s published USD API rates. Do not quote them as CAD. CAD is on the invoice.
| Model | Input / 1M | Output / 1M | Cache read / write |
|---|---|---|---|
| Fable 5.1 | $10 | $50 | $0.25 / $12.50 |
| Opus 5 | $5 | $25 | $0.50 / $6.25 |
| Sonnet 5 | $2 | $10 | $0.20 / $2.50 |
| Haiku 4.5 | $1 | $5 | $0.10 / $1.25 |
Anthropic says Fable 5.1 cache reads now cost $0.25 per million tokens, 75% less than Fable 5. The legacy Fable 5 card still lists cache reads at $1 per million. Anthropic estimates that change reduces the cost of typical workloads by about 25% and highly agentic workloads by up to about 45%. Those are Anthropic’s estimates, not this desk’s measurements.
Batch processing is listed at 50% off. Opus 5 Fast mode is listed at 2x standard pricing for about 2.5x faster speeds. US-only inference is 1.1x on input and output tokens. That is extra USD, not a Canadian privacy feature.
Claude Mythos 5.1 is Anthropic’s newest Mythos-class model, aimed at cybersecurity and biology. Access is limited to a small set of vetted organizations. Pricing starts at $10 / $50. Anthropic says it is currently available to a set of US organizations. It is not the Alberta shop SKU. Do not order it as the default automation path.
What Alberta operators should do
Fable 5.1 is the expensive long-horizon SKU. Most Alberta automation still belongs on Sonnet 5 at $2 / $10 unless the job actually needs hours-long agents. If a Sherwood Park or Edmonton shop is paying Fable prices for inbox triage, that is the wrong card.
Default 30-day data retention for safety monitoring is not a Canadian residency story. Eligible enterprise customers can store data on their own cloud under Enterprise Frontier Safeguards; until EFS is available, Anthropic says eligible customers can use Fable 5.1 with zero data retention. That is the private-AI angle. Confirm eligibility on Anthropic’s paper. Do not treat a Pro seat as EFS.
Cursor shops: Fable inside Cursor is still cloud Anthropic. Keep client files, tenders, and credentials on a private AI security path. A model upgrade is not a data-residency policy. If the workflow has to stay on the desk, that is AI consulting work, not a public paste.
Safeguards matter on the invoice. Anthropic says many cybersecurity and biology queries may be routed to less capable models, and you will not be charged Fable prices for rerouted requests.
The guardrail
Price is not permission. A $10 / $50 card does not put inference in Alberta. US-only 1.1x is a location surcharge, not a residency win. For AI automation Alberta work, pick the cheapest model that actually finishes the job, then keep the files that would hurt if they leaked off the public prompt.