Claude Opus 5.5 vs Fable 5.2: API price and when to use which

Claude Opus 5.5 is Anthropic’s recommended starting model for most coding and knowledge work. Claude Fable 5.2 is for more demanding, long-horizon reasoning. This page answers both “Opus 5.5 vs Fable 5.2” and “Fable 5.2 vs Opus 5.5”. On Code By AI both models use the same prepaid key and the same published price structure: Opus 5.5 (`claude-opus-5-5`) is billed at the published 1/4 of Anthropic’s list price ($1.00 input / $5.00 output per 1M tokens); Fable 5.2 (`claude-fable-5-2`) is billed at the published 1/4 ($2.00 / $10.00). There is no monthly subscription.

Short answer

Start with Claude Opus 5.5 for most coding and knowledge work, following Anthropic’s guidance, and consider Fable 5.2 for demanding long-horizon work; current Code By AI catalog prices per 1M input / output tokens are Claude Opus 5.5 $1.00 / $5.00 and Claude Fable 5.2 $2.00 / $10.00 in USD.

Token and monthly figures are arithmetic from the current catalog, not a forecast or a measured benchmark. The monthly example excludes cache writes, tools and actual usage differences.

Token prices on the same basis

All values are USD per 1M tokens. Code By AI prices already include the published multiplier. Vendor list prices are shown only when verified. The 5-minute and 1-hour cache writes apply only to Claude and use each platform’s own input rate.

ModelMetricCode By AIOfficial listOpenRouter
Claude Opus 5.5claude-opus-5-5Input$1.00$4.00$4.00
Cache read$0.05$0.20$0.20
Cache write (5 min)$1.25$5.00$5.00
Cache write (1 hour)$2.00$8.00$8.00
Output$5.00$20.00$20.00
Claude Fable 5.2claude-fable-5-2Input$2.00$8.00Not verified
Cache read$0.05$0.20Not verified
Cache write (5 min)$2.50$10.00Not verified
Cache write (1 hour)$4.00$16.00Not verified
Output$10.00$40.00Not verified

OpenRouter standard endpoint (verified): anthropic/claude-opus-5.5 · anthropic · anthropic/claude-opus-5.5-20260921 ·

OpenRouter cells show a number only when the exact provider model ID and rate basis were checked against this catalog row. Cells marked Not verified have no such evidence. Credit-purchase fees and other OpenRouter plans are separate and are never added to the per-token prices here. OpenRouter pricing and FAQ · FAQ

Unverified rates are not inferred from catalog values. /pricing

Which model should I start with?

Start with Claude Opus 5.5 for most coding and knowledge work; switch to Claude Fable 5.2 when the task is long-horizon or official guidance says Opus at higher effort is still not enough.

That recommendation comes from Anthropic, not from a Code By AI evaluation. This page does not publish third-party score tables or any self-measured speed or quality figures. On Code By AI you do not need a new key to switch: keep the same prepaid key and base URL, and change only the model ID.

Anthropic’s models overview says:

If you're unsure which model to use, start with Claude Opus 5.5 for most workloads. Use Claude Fable 5.2 for demanding reasoning and long-horizon agentic work, or when your evals on Claude Opus 5.5 at higher effort still fall short.

Anthropic models overview

The Claude Fable 5.2 announcement adds:

Fable and Mythos are best for the most ambitious, open-ended technical work. For most coding and knowledge tasks, we recommend that users start with Claude Opus 5.5, our strongest daily driver, and then switch to a different model as required, depending on the scope and complexity of their work—upward to Fable or Mythos, or downward to Claude Sonnet 5.5 or Claude Haiku 5.5. Fable 5.2 can also play the role of Advisor (where Haiku, Sonnet, or Opus models consult it to help solve harder problems) or Orchestrator (where it delegates simpler tasks to smaller models).

Claude Fable 5.2 and Mythos 5.2 announcement

Open the Code By AI model pages for each ID:

How do official capabilities differ?

Anthropic positions Claude Opus 5.5 for long-running agentic coding and knowledge work, and Claude Fable 5.2 for demanding reasoning and long-horizon agentic work. The official differences below are quoted from Anthropic; they are not Code By AI measurements.

Claude Opus 5.5 is built for long-running agentic coding and knowledge work. Typical uses include multihour autonomous coding agents, large-scale refactoring, complex systems engineering, vision-heavy workflows, and computer use.

Claude Opus 5.5 overview

Claude Fable 5.2 is built for demanding reasoning and long-horizon agentic work. Typical uses include agent sessions that run for hours, multistep deep research, and analysis carried through to a finished document, spreadsheet, or deck.

Claude Fable 5.2 overview

Default effort on the Claude API

On Anthropic’s Claude API, the default effort parameter is medium for Claude Opus 5.5 and high for Claude Fable 5.2. Anthropic describes this as the default on the official API; set effort explicitly to use a different level.

Claude Opus 5.5Claude Fable 5.2
mediumhigh

According to Anthropic’s model overviews, both models have a 1M context window, 128K max synchronous output, a June 2026 knowledge cutoff, always-on adaptive thinking, and moderate comparative latency (a relative band in the official lineup, not a measured speed).

Official Claude API rules (not Code By AI product claims)

On the Message Batches API, Claude Opus 5.5 supports up to 300k output tokens with the output-300k-2026-03-24 beta header. Anthropic’s models overview lists that 300k Message Batches output beta for Claude Opus 5.5 and does not list Claude Fable 5.2. This is a rule of Anthropic’s official API. This page does not claim that Code By AI offers the Message Batches API or that 300k-output beta. Anthropic models overview · Claude Opus 5.5 overview

Claude Fable 5.2 requires 30-day data retention and isn't available under zero data retention unless expressly authorized by Anthropic. This is a requirement of Anthropic’s official API, not a statement of Code By AI’s data-retention policy. Claude Fable 5.2 overview · Anthropic API data retention

What kind of task should I pick?

The rows below restate official typical uses and official selection guidance. They are not a Code By AI benchmark or a ranked quality list.

Kind of workOfficial starting pointSource
Most coding and knowledge tasksClaude Opus 5.5Models overview and Fable 5.2 announcement
Multihour autonomous coding agents, large-scale refactoring, complex systems engineering, vision-heavy workflows, and computer useClaude Opus 5.5Opus 5.5 overview (typical uses)
Demanding reasoning and long-horizon agentic work, or when evals on Opus 5.5 at higher effort still fall shortClaude Fable 5.2Models overview
Agent sessions that run for hours, multistep deep research, and analysis carried through to a finished document, spreadsheet, or deckClaude Fable 5.2Fable 5.2 overview (typical uses)
The most ambitious, open-ended technical work, or using Fable as an Advisor or Orchestrator alongside other Claude modelsClaude Fable 5.2Fable 5.2 announcement

Official sources

What does each model cost on Code By AI?

On Code By AI both models use the same prepaid key. The table shows the final USD token prices; no additional conversion is needed. Each model’s price is based on its official list price and the platform rate shown below.

Opus 5.5 · Anthropic listOpus 5.5 · Code By AIFable 5.2 · Anthropic listFable 5.2 · Code By AI
Input$4.00$1.00$8.00$2.00
Cache read$0.20$0.05$0.20$0.05
Cache write (5-minute)$5.00$1.25$10.00$2.50
Cache write (1-hour)$8.00$2.00$16.00$4.00
Output$20.00$5.00$40.00$10.00

USD per 1M tokens. Code By AI prices already include the published multiplier and ÷6 factor; do not divide them again.

The wallet and the published model prices use the same USD unit. Before discounts and fees, a $10 top-up adds $10 to the wallet.

Anthropic’s official Fast mode list prices are $8.00 input / $40.00 output for Opus 5.5 and $16.00 / $80.00 for Fable 5.2, per 1M tokens. This page does not state a Code By AI Fast mode price for either model. Anthropic API pricing

See each model page and the model price documentation for the same catalog rates and a worked usage example.

How do I call each model on Code By AI?

In the same terminal, replace sk-proj-xxxx with your API key and run the setup below, then run either Messages request. Claude Code and both curl examples use the same ANTHROPIC_AUTH_TOKEN.

The default Claude Code model on Code By AI is claude-sonnet-5, not these two. Omit the model ID and the request uses that default; to call Opus 5.5 or Fable 5.2 you must pass the ID explicitly.

Shared setup for Claude Code and curl

export ANTHROPIC_BASE_URL="https://www.codebyai.net/api/agent"
export ANTHROPIC_AUTH_TOKEN="sk-proj-xxxx"

Minimal Messages request for claude-opus-5-5

curl --fail-with-body --max-time 120 https://www.codebyai.net/api/agent/v1/messages \
  -H "x-api-key: ${ANTHROPIC_AUTH_TOKEN:?Set ANTHROPIC_AUTH_TOKEN}" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-opus-5-5","max_tokens":256,"messages":[{"role":"user","content":"Hello"}]}'

Minimal Messages request for claude-fable-5-2

curl --fail-with-body --max-time 120 https://www.codebyai.net/api/agent/v1/messages \
  -H "x-api-key: ${ANTHROPIC_AUTH_TOKEN:?Set ANTHROPIC_AUTH_TOKEN}" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-fable-5-2","max_tokens":256,"messages":[{"role":"user","content":"Hello"}]}'
Claude Code setup guide

FAQ

Which should I pick, Opus 5.5 or Fable 5.2?

Start with Claude Opus 5.5 for most coding and knowledge work; use Claude Fable 5.2 when you need the heavier long-horizon model, following Anthropic’s guidance.

Does this page also cover “Fable 5.2 vs Opus 5.5”?

Yes. This single URL answers both “Opus 5.5 vs Fable 5.2” and “Fable 5.2 vs Opus 5.5”.

What are the model IDs on Code By AI?

The model IDs are `claude-opus-5-5` and `claude-fable-5-2`. They match the IDs on the Code By AI model pages.

What do they cost on Code By AI?

Opus 5.5 is billed at the published 1/4 of the official list price ($1.00 input / $0.05 cache read / $5.00 output per 1M tokens). Fable 5.2 is billed at the published 1/4 ($2.00 / $0.05 / $10.00). Those published prices already include the model multiplier and the documented ÷6 factor.

Do both models use the same key on Code By AI?

Yes. Both models are called with the same prepaid key, the same base URL, and the same published price structure. Change only the model ID.

Do I need a new key to switch between them?

No. Change the model ID on the same prepaid key. The default Claude Code model is not these two, so the ID must be set explicitly.

How does top-up work?

There is no subscription. Before discounts and fees, $10 adds $10 to the wallet; usage is debited at the published token prices on this page and in Model Plaza.

Are the context windows different?

No. Both models have a 1M context window and 128K max synchronous output, according to Anthropic’s model overviews.

Is Mythos available on Code By AI?

No. Code By AI does not offer Mythos.

How do I use them in Claude Code?

Set ANTHROPIC_BASE_URL to https://www.codebyai.net/api/agent and ANTHROPIC_AUTH_TOKEN to your key, as in the Claude Code setup guide, then pass the model ID.

How do Code By AI prices compare with official list prices?

Code By AI prices are 1/4 of the official list price for claude-opus-5-5 and 1/4 for claude-fable-5-2. The final per-token price also depends on each model’s official list price. The table above shows the final USD rates; no additional conversion is needed.

Does Code By AI provide Anthropic’s Message Batches 300K output beta or Fable’s 30-day retention rule?

No. The 300K Message Batches output beta and the 30-day data-retention requirement are rules of Anthropic’s official API, quoted above. This page does not claim that Code By AI offers the Message Batches API, and it does not state Code By AI’s data-retention policy.

Can I use another language’s docs to set this up?

Yes. This compare page is published in English, Traditional Chinese, Korean, French, Japanese, Russian, Spanish, Vietnamese, Indonesian and Portuguese. Use the language selector; documentation coverage varies by page.

Your next step

Create a Code By AI account, top up the prepaid wallet, and create an API key. Then call Opus 5.5 or Fable 5.2 with the snippets above, or follow the Claude Code setup guide.

Code By AI is a third-party prepaid API gateway. It is not an official Anthropic product.