Claude Fable 5.1 Complete Guide — What Does a 75% Drop in Cache Pricing Actually Change?
Anthropic released Claude Fable 5.1 on September 1, 2026, cutting prompt cache read pricing by 75% ($1.00 → $0.25 per million tokens). This translates to roughly 25% cost savings on typical tasks and up to 45% on complex coding. With a 1M-token context window and an intelligence benchmark score of 53, it's the go-to model for long documents and extended agentic tasks.
"I'd love to use AI, but I'm worried about the cost." If that's been holding you back, here's good news. On September 1, 2026, Anthropic released Claude Fable 5.1 — performance stays the same, but the pricing that heavy users care about most has dropped significantly.
What Is It?
This is the top-tier model in Anthropic's Claude AI assistant lineup. It excels at complex tasks and processing large volumes of text.
What Changed
The biggest change is a 75% cut in the cache read price, effective September 1, 2026.
Fable 5 (Before) Cache reads: $1.00 per million tokens. Tasks that repeatedly reference the same material — like long documents — would rack up costs fast.
Fable 5.1 (After) Cache reads: $0.25 per million tokens — one quarter of the previous price. The more you reuse the same material, the greater the savings.
The Numbers at a Glance (as of September 1, 2026)
| Item | Fable 5 | Fable 5.1 |
|---|---|---|
| Input tokens (per million) | $10 | $10 (unchanged) |
| Output tokens (per million) | $50 | $50 (unchanged) |
| Cache reads (per million) | $1.00 | $0.25 |
| Cost savings on typical tasks | — | ~25% reduction |
| Cost savings on complex coding | — | up to 45% reduction |
Cache reads refer to the system where the AI "remembers" documents it has already read. When you repeatedly reference the same long contract, manual, or codebase, the AI doesn't have to re-read from scratch each time — and that reuse is now 75% cheaper. That's the core of this update.
Performance improvements are also worth noting. The context window (the amount of text the model can process in one go) is 1 million tokens. The overall intelligence benchmark score is 53 — well ahead of Claude Sonnet 5's score of 38. On agentic benchmarks (tasks requiring multi-step autonomous action), it achieves more than twice the score of Fable 5.
How Can You Actually Use It?
Scenario 1: Reviewing a thick contract (for legal or administrative teams)
Imagine walking a 100-page English-language contract through Claude multiple times, verifying it clause by clause. With Fable 5.1's lower cache read pricing, repeatedly referencing the same document costs significantly less. You can ask, "Does this clause conflict with Japanese law?" or "Identify the risks compared to a similar deal" as many times as you need.
Scenario 2: Supporting long-term research (for researchers and planners)
Feed in market research reports or academic papers, then ask questions, organize findings, and request summaries dozens of times. The material stays cached, so you don't have to reload it each time. Anthropic estimates this kind of high-agentic task sees up to 45% cost savings.
Scenario 3: Refactoring complex code (for engineers)
Load a large codebase and iteratively fix bugs, add features, and write tests. Fable 5.1's maximum output is 128,000 tokens, allowing large amounts of code to be generated in a single instruction. Its agentic benchmark score is 80.2 — about 22% higher than Sonnet 5's 65.9 — making it reliably strong for multi-step workflows.
For routine, non-complex tasks — like writing short emails, quick translations, or simple summaries — Claude Sonnet 5 is faster and more cost-efficient. Think of Fable 5.1 as your "heavy-duty" model for tasks that really need it, and use both accordingly.
What Does It Cost?
- ✓Processing speed: 61.0 tokens/sec
- ✓Best for routine and short-text tasks
- ✓Great for cost-sensitive projects
- ✓Agentic score: 65.9
- ✓Processing speed: 52.8 tokens/sec
- ✓Best for complex reasoning and long-term tasks
- ✓Cache reads now 75% cheaper
- ✓Agentic score: 80.2
Available Platforms (as of September 1, 2026)
- Claude API (direct access)
- Amazon Bedrock
- Google Cloud
- Microsoft Foundry
Claude Mythos 5.1, released alongside Fable 5.1, is a limited-access model for cybersecurity and life sciences use cases. It is not available to general users. For everyday work, Fable 5.1 is the model you'll use.
Summary
Claude Fable 5.1 has been available since September 1, 2026, with cache read pricing cut by 75%. Input and output token prices remain unchanged. It's especially recommended for anyone who frequently references long documents during work or wants to delegate complex, multi-step tasks to AI. For shorter or more routine work, Sonnet 5 is sufficient. Start by figuring out which pattern fits your workflow, then choose accordingly.
Related Articles
Check out these related articles as well.
More Claude articles
Which Free AI Is Smartest? Claude vs ChatGPT vs Gemini (October 2026)
October 2026 reshuffled the free tiers of the big three AI assistants. Claude's free plan now runs the latest Sonnet 5.5, while Gemini's free tier is limited to a lightweight model. Here's how the free and ~$20/month plans compare, so you can pick the right one.
Claude Sonnet 5.5 Is Free to Use — 30% Faster Replies and Near-Flagship Smarts
Anthropic released Claude Sonnet 5.5 on September 28, 2026. It responds more than 30% faster than Sonnet 5 and nearly matches the top-tier Opus 5.5 on office work, and it's available on the free plan, so anyone can try a high-end AI at no cost.
What Is an AI Agent? How It Differs from a Chatbot, Explained Simply
An AI agent is an AI that doesn't just answer questions but actually does the work for you. This guide explains the difference from chatbots with everyday analogies, covers agents you can use in 2026 like ChatGPT agent mode and Claude Cowork, what they can and can't do, and how to use them safely.