AnthropicTextVision

claude-fable-5

claude-fable-5 is Anthropic's model at $10.00 input and $50.00 output per million tokens, with a 1049K context window. On a typical 3:1 input-to-output mix that blends to roughly $20.00 per million tokens, placing it in the premium band of the 89 models tracked here — cheaper than 0% of them.

Open cost calculator View provider Open official platform

MODEL OVERVIEW

Context window1049K

Input price$10 / 1M

Output price$50 / 1M

Output speed60 tok/s

Quality score96

Data updatedUpdated Jun 2026

Price verified2026-07-11 · Official

By price, claude-fable-5 ranks 85th cheapest of 89 on input and 86th on output. Its quality score of 96 ranks 6th of 89, and on value (quality score ÷ output price) it ranks 78th of 89. Its blended rate sits 549% above the catalog average.

In real billing terms: a support chatbot handling 50,000 turns a month (2,000 input / 500 output tokens each) costs about $2,250 per month on claude-fable-5, while a RAG workload of 20,000 queries (8,000 input / 800 output tokens each) runs about $2,400. Because output is priced 5.0× input here, long-generation tasks scale cost fastest — cap output length before you optimise anything else.

A 1049K context window holds roughly 1,399 pages of A4 text (at ~750 tokens per page), ranking 8th of 89 here. That is enough to drop in an entire technical manual or several source files at once, which suits long-document summarization and codebase-wide analysis. Output runs at about 60 tok/s (67th fastest of 89), acceptable interactively, though long responses will feel slow.

We track 12 Anthropic models, and claude-fable-5's blended rate sits 110% above that provider's average. If your task tolerates a little less headroom, deepseek-v4-pro (DeepSeek) blends about 95% cheaper at a quality score of 93, and is worth benchmarking first. Going the other way, gpt-5.6-sol scores 98 on quality for about 60% less.

The closest model in Anthropic's line-up is claude-fable-5.1 ($10.00 input / $50.00 output, quality 97). Against it, claude-fable-5 scores 1 point lower on quality, runs 6 tok/s slower (about 9%), has a 49K larger context window. When two models sit this close on price, the deciding factor is usually the quality ceiling — run both against your real prompts before picking.

Our read on it

Twice the price of Opus 5 and one point lower on quality. This page exists to explain why that can still be rational.

Start with what looks indefensible

claude-fable-5 is $10/$50 at quality 96. claude-opus-5 is $5/$25 at quality 97. Same vendor, same 1,049K context, same text and vision modalities, effectively the same speed (60 against 58 tokens/second). Double the price, and not one column in its favour.

Anyone deciding from a comparison table should stop here and pick Opus 5, and that is usually correct. We are not going to launder the gap with marketing language.

The dimension the score cannot measure

Our quality score is a composite benchmark figure. It tracks the kind of thing an automated grader can check: task accuracy, instruction following, factual consistency. The fable line is positioned around writing and sustained narrative — tonal consistency, a character voice that holds, structure that survives tens of thousands of words. Those are precisely the properties automated grading is worst at.

So the honest statement is this: one point lower does not mean the prose is one point worse, and it does not mean the prose is better either. It is a difference you can only settle with a blind test on your own material, not one you can read off a table. Worth noting too that claude-mythos-5 sits at the identical $10/$50 and quality 96 — the two differ in intended use, fable tagged for coding and reasoning, mythos for chat and reasoning.

Who should actually pay the premium

The window is narrow: teams producing a high volume of prose that reaches a reader more or less as generated. Publishing, screenwriting, long-form marketing, game narrative. In those settings a draft that needs a heavy editorial pass costs more in labour than a month of API billing, so the $50 output rate is not a cost centre — it is leverage against editor hours.

If your output gets templated, forced into a fixed structure, or runs to a few hundred words, the premium buys nothing. The same applies to code: fable carries the coding tag, but Opus 5 scores higher for half the price. The pragmatic route is to baseline on Opus 5 and only run a blind A/B on real material once you actually hit the problem where the text reads wrong.

Estimated monthly cost

Monthly API spend across four common workloads, next to the cheapest model we track.

Workload	Usage assumption	This model	Cheapest model
Support chatbot	2,000 input / 500 output tokens per turn, 50,000 turns per month	$2,250	$3.00
RAG document Q&A	8,000 input / 800 output tokens per query, 20,000 queries per month	$2,400	$3.84
Coding assistant	12,000 input / 2,000 output tokens per request, 10,000 requests per month	$2,200	$3.20
Batch summarization	4,000 input / 400 output tokens per item, 200,000 items per month	$12,000	$19

Derived directly from the rates on this page. Excludes prompt caching, batch discounts and free tiers.

How it compares

The alternatives worth benchmarking alongside it, from the same dataset.

Model	Provider	Blended /1M	Quality	Price delta
claude-fable-5	Anthropic	$20.00	96	—
deepseek-v4-proCheaper option	DeepSeek	$0.99	93	-95%
gpt-5.6-solHigher quality	OpenAI	$8.00	98	-60%
claude-fable-5.1Same provider	Anthropic	$20.00	97	+0%
qwen3.5Best value overall	Alibaba	$0.11	91	-99%

Best for

CodingReasoning

Modalities

TextVision

Where it fits — and where it doesn't

Good fit

Coding: writing, refactoring and debugging across multi-file project context.
Reasoning: multi-step decomposition, planning and tasks that need rigorous inference.
Long documents: a 1049K window ingests a full contract, paper or large codebase in one pass.
Multimodal input: handles images, screenshots and scanned documents directly.

Poor fit

High-frequency batch jobs: at $20.00 blended per million tokens, bills scale quickly at volume.

Frequently asked questions

How much does the claude-fable-5 API cost?

claude-fable-5 costs $10.00 per million input tokens and $50.00 per million output tokens. On a 3:1 input-to-output mix that blends to about $20.00 per million tokens.

What does claude-fable-5 cost per month?

It depends on volume. Using the four worked scenarios on this page: a support chatbot runs about $2,250/month, RAG document Q&A about $2,400/month, a coding assistant about $2,200/month, and batch summarization about $12,000/month. Plug your own token counts into the cost calculator for a tailored figure.

How large is the claude-fable-5 context window?

1049K tokens — roughly 1,399 pages of A4 text at ~750 tokens per page, ranking 8th of the 89 models tracked here.

claude-fable-5 or deepseek-v4-pro — which is better value?

deepseek-v4-pro blends to about $0.99 per million tokens versus $20.00 for claude-fable-5 — roughly 95% cheaper — at a quality score of 93 against 96. If that quality gap does not show up on your task, take the cheaper one; if it does, stay with claude-fable-5. Benchmark both on your own data before deciding.

What is claude-fable-5 best suited for?

Tagged use cases: coding, reasoning. Coding: writing, refactoring and debugging across multi-file project context. Less suitable: High-frequency batch jobs: at $20.00 blended per million tokens, bills scale quickly at volume.

How current is this claude-fable-5 pricing?

Pricing here was last verified on 2026-07-11 against official documentation. llmprice.app runs a daily collection job, but providers can change rates between runs — confirm against Anthropic's official pricing page before you commit.