Breaking

Anthropic Releases Claude Fable 5.1 With Five API Optimization Levers

September 2, 2026 at 07:18 EDT

Foundation Models
AI Agents
Software Dev & Coding

Anthropic released Claude Fable 5.1 on September 1 with lower cache-read pricing and new controls for balancing reasoning quality, latency, and cost. The model is generally available through the Claude API and major cloud platforms, while its practical optimization playbook gives developers five ways to tune existing applications immediately.

September 1, 2026 · Anthropic

Five levers make Fable 5.1 cheaper to run

Claude Fable 5.1 combines frontier-level long-horizon work with lower cache-read costs and practical controls for tuning effort, prompts, and migrations.

1MContext window

128KMaximum output

25%Typical cost reduction

45%Maximum estimated agentic savings

Cache reads cost one quarter as much

The input and output rates stay the same; the largest pricing change is repeated context.

$1.00

Claude Fable 5

$0.25

Claude Fable 5.1

One blue block represents one quarter of the previous cache-read price.

A practical optimization flow

Tune the model around the work each turn actually requires.

Set effortUse low for routine work

→

Run diagnosticsProfile cost and cache use

→

Audit promptsRemove dated scaffolding

→

Migrate configVerify tools and model IDs

Where it helps

Long-running coding, research, code review, incident analysis, and unattended prototypes can benefit from stronger reasoning with controlled effort.

What to watch

Low effort may search less often, forced tool choice returns an error, and changing configuration incorrectly can lose cache hits.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.