Claude Fable 5

anthropic/claude-fable-5
NouveauEn vedette
par Anthropic · text+image+file in · text out · 1M ctx · 2026-06-09

Anthropic's Claude Fable 5: 1M-context model scoring 85.0 on OSWorld-Verified, accessed via OrcaRout…

Points de terminaison:/v1/messages/v1/chat/completions
ENTRÉE$10.00/ 1M tokens
SORTIE$50.00/ 1M tokens
p50 TTFT5.99 s7 j
TTFT p9537.71 s7 j
TRAFIC1.8Mtokens / 7 j

What is Claude Fable 5 and who should use it?

Claude Fable 5 is a large language model from Anthropic that emphasizes extended context and multimodal input. Its 1,000,000-token context window allows a single request to include entire books, extensive codebases, or long conversation histories. The model accepts text, images, and file uploads, and generates up to 128,000 tokens per response. Pricing is transparent: $10.00 per million input tokens and $50.00 per million output tokens, passed through from Anthropic with no additional markup by OrcaRouter. This model is appropriate for teams working on high-stakes document review, complex analysis, or advanced agentic systems. Because of its premium pricing, it should be reserved for tasks that leverage its exceptional context length and multimodal capabilities, rather than for routine or low-cost operations.

What input formats does Claude Fable 5 support?

Claude Fable 5 accepts three input modalities: text, image, and file. Text can be provided as plain strings or structured messages. Image inputs support common formats such as JPEG, PNG, GIF, and WebP, and the model can reason about visual content including diagrams, charts, and natural scenes. File input allows uploading documents (e.g., PDF, Word, or plain text) which are processed as part of the context. Together, these modalities enable tasks like extracting information from scanned documents, analyzing figures, or processing mixed-media reports. All inputs are counted against the 1,000,000-token limit. For images, token consumption depends on image size and detail level, following Anthropic's tokenization rules. This flexibility makes the model suitable for domains like healthcare imaging, legal case files, and technical documentation.

How does the large context window affect performance?

The 1,000,000-token context window allows Claude Fable 5 to process entire datasets or long-running conversations without chunking or summarization. In practice, this means the model can maintain coherence over extremely long sequences, which is critical for tasks like analyzing research literature, auditing long code files, or simulating dialogues over many turns. However, longer contexts increase inference latency and token costs. The model outputs up to 128,000 tokens per request, enabling detailed responses. While the large context is a differentiator, users should be aware that not all use cases need the full window; for shorter inputs, a smaller model may offer faster and cheaper performance. On OSWorld-Verified, the model scored 85.0, suggesting strong capability in tasks that require sustained attention and gradual reasoning.

Exemples de code

from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="$ORCAROUTER_API_KEY",
)

response = client.chat.completions.create(
    model="anthropic/claude-fable-5",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Tarifs

Entrée / 1M tokens$10.00
Sortie / 1M tokens$50.00
Lecture cache / 1M$1.00
Écriture cache / 1M$12.50
DeviseUSD

Performances

p50 TTFT
5.99 s
Vitesse de sortie
81.9 tok/s
TTFT p95
37.71 s
Taux d'erreur
0.61%

Benchmarks publics

AutomationBench
17.4
Blueprint-Bench 2
38.6
FrontierCode (Diamond)
29.3
GDPval.pdf
29.8
Legal Agent Benchmark
13.3
OSWorld-Verified
85.0
SWE-Bench Pro
80.3
Source: artificialanalysis.ai

FAQ

What is the cost per token for Claude Fable 5 on OrcaRouter?
The cost is $10.00 per 1 million input tokens and $50.00 per 1 million output tokens, billed at Anthropic's provider rate with zero markup from OrcaRouter.
What is the context window size of Claude Fable 5?
The context window is 1,000,000 tokens (1M tokens). The maximum output is 128,000 tokens per request.
What are the main strengths of Claude Fable 5?
Its main strengths are an extremely large context window (1M tokens), multimodal input (text, image, file), a high score of 85.0 on the OSWorld-Verified agentic benchmark, and a 128K output limit.
How does Claude Fable 5 compare to Claude 3 Opus?
Fable 5 has a larger context window (1M vs 200K), higher OSWorld score (85.0 vs ~55-60), larger output limit (128K vs 4,096), and is cheaper per token ($10/$50 vs $15/$75). Opus may still be suitable for shorter tasks at lower absolute cost.
Does OrcaRouter add any markup or hidden fees for Claude Fable 5?
No, OrcaRouter adds zero markup. You pay exactly $10.00 per 1M input tokens and $50.00 per 1M output tokens, with no additional per-request or subscription fees.
How do I call Claude Fable 5 via OrcaRouter's OpenAI-compatible API?
Use base URL https://api.orcarouter.ai/v1 and model ID "anthropic/claude-fable-5". Send a POST request to /chat/completions with standard OpenAI parameters. Include your OrcaRouter API key in the Authorization header.
What input modalities does Claude Fable 5 support?
It supports text, image (JPEG, PNG, GIF, WebP), and file uploads (PDF, Word, plain text, etc.). It does not support audio or video.
Is Claude Fable 5 suitable for real-time applications?
It can be used for real-time applications if latency is tolerable, but with large contexts it may be slow. For low-latency needs, consider smaller models like Claude 3 Haiku or GPT-4o mini. Streaming is supported.
How is data handled when using Claude Fable 5 through OrcaRouter?
OrcaRouter passes requests directly to Anthropic's API. Anthropic's privacy policy governs data handling. OrcaRouter may log metadata (e.g., timestamps, token counts) for billing and monitoring but does not store prompt contents beyond what is necessary for API operation. Check OrcaRouter's privacy policy and Anthropic's data policy for details.
What if I need a smaller context or lower cost? Can OrcaRouter help me switch models?
Yes, OrcaRouter's API allows you to dynamically change the model ID per request. You can use Claude Fable 5 for complex tasks and a cheaper model (e.g., Claude 3 Sonnet or GPT-4o mini) for simpler ones within the same codebase.

Intégrer ce badge

Anthropic: Claude Fable 5$10.00/M in5985ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/anthropic/claude-fable-5" target="_blank"> <img src="https://www.orcarouter.ai/embed/anthropic/claude-fable-5.svg" alt="Anthropic: Claude Fable 5 sur OrcaRouter" /> </a>
Markdown [![Anthropic: Claude Fable 5](https://www.orcarouter.ai/embed/anthropic/claude-fable-5.svg)](https://www.orcarouter.ai/models/anthropic/claude-fable-5)