Claude Fable 5.2 API
claude-fable-5-2Claude Fable 5.2 is Anthropic's most capable widely released model, released on 9 November 2026 as the successor to Fable 5.1, and on APIMODELS it costs $4.80 per 1M input tokens and $24.00 per 1M output — 40% below Anthropic's $8 / $40 list — with prompt-cache reads at $0.12 and cache writes at $6.00 for both the 5-minute and the 1-hour TTL (Anthropic charges $0.20, $10 and $16). Anthropic cut the list price by 20% from Fable 5.1's $10 / $50 and says outputs are up to 30% faster. On Anthropic's launch table it leads both Fable 5.1 and Opus 5.5 on agentic coding — FrontierCode 2.0 49.1% at xhigh effort (Fable 5.1 32.4%, Opus 5.5 45.5%) and FrontierSWE v2 69.9% (56.3%, 62.3%) — and on long-horizon and knowledge work: Terminal-Bench-Science 0.1 74.6% (52.6%, 58.7%), Remote Labor Index 41.3% (17.9%, 21.3%), Humanity's Last Exam with tools 74.2% (65.6%, 67.7%) and Chartography without tools 73.4% (44.8%, 64.4%). It is not first everywhere: Gemini 4 Carbon scores higher on AutomationBench (53.9% against 51.4%). Anthropic also reports that at high effort Fable 5.2 matches Opus 5.5's best FrontierCode score for roughly half the cost per task; both models are 40% below list here, so that ratio holds on APIMODELS too. Specifications: a 1M-token context window, 128K max output, and text and image input. Thinking is always on — control its depth with output_config.effort (low to max, default high). temperature, top_p, top_k and budget_tokens are removed on this model upstream, and APIMODELS drops them instead of forwarding them. Anthropic rejects forced tool use (tool_choice any or a named tool) on Fable 5.2; APIMODELS sends it as tool_choice auto plus an instruction to call that tool, which is also how structured outputs run here. Assistant prefill is not supported. Anthropic's safeguards decline most high-risk cybersecurity work, such as penetration testing and exploit generation, and those requests come back with stop_reason refusal. Call it through /v1/messages (Claude Code, the Anthropic SDK, Cursor) or the OpenAI-compatible /v1/chat/completions with one key; failed requests are not charged.
API Documentation
View complete API reference with all parameters and examples.
Advanced Features
Streaming
Enable real-time streaming responses with Server-Sent Events.
{
"model": "claude-fable-5-2",
"stream": true,
"max_tokens": 1024,
"messages": [...]
}Function Calling (Tools)
Enable Claude to use tools and call functions.
{
"model": "claude-fable-5-2",
"max_tokens": 1024,
"tools": [{
"name": "get_weather",
"description": "Get current weather for a location",
"input_schema": {
"type": "object",
"properties": {
"location": {"type": "string", "description": "City name"}
},
"required": ["location"]
}
}],
"tool_choice": {"type": "auto"},
"messages": [{"role": "user", "content": "What's the weather in Tokyo?"}]
}PDF Support
Analyze PDF documents by sending them as base64 encoded content.
{
"model": "claude-fable-5-2",
"max_tokens": 1024,
"messages": [{
"role": "user",
"content": [{
"type": "document",
"source": {
"type": "base64",
"media_type": "application/pdf",
"data": "<base64_encoded_pdf>"
}
}, {
"type": "text",
"text": "Summarize this document."
}]
}]
}Structured Output (JSON Schema)
Get structured JSON responses that match your schema.
{
"model": "claude-fable-5-2",
"max_tokens": 1024,
"output_format": {
"type": "json_schema",
"schema": {
"type": "object",
"properties": {
"name": {"type": "string"},
"age": {"type": "integer"}
},
"required": ["name", "age"]
}
},
"messages": [{"role": "user", "content": "Extract info: John is 30 years old"}]
}Web Search
Enable Claude to search the web for up-to-date information.
{
"model": "claude-fable-5-2",
"max_tokens": 1024,
"tools": [{
"type": "web_search_20250305",
"name": "web_search",
"max_uses": 5
}],
"messages": [{"role": "user", "content": "What's the latest news about AI?"}]
}API Parameters Reference
| Parameter | Type | Required | Description |
|---|---|---|---|
| model | string | Yes | Model identifier (e.g., claude-fable-5-2) |
| messages | array | Yes | Array of message objects with role and content |
| max_tokens | integer | Yes | Maximum tokens in the response (1 - 128000) |
| system | string | No | System prompt to set context |
| stream | boolean | No | Enable streaming responses (SSE) |
| temperature | number | No | Sampling temperature (0.0 - 1.0) |
| top_p | number | No | Nucleus sampling threshold (0.0 - 1.0) |
| top_k | integer | No | Top-k sampling (0 - infinity) |
| stop_sequences | array | No | Sequences that stop generation |
| tools | array | No | Function calling tools definition |
| tool_choice | object | No | Tool selection strategy (auto/any/tool) |
| thinking | object | No | Enable extended thinking mode |
| output_format | object | No | Structured output with JSON schema |
Full API Documentation
View complete API reference with streaming, thinking, and more.
Pricing
Billing: Cost = (input_tokens * input_price + output_tokens * output_price) / 1,000,000
40% Below Official
$4.80 / $24.00 per 1M tokens against Anthropic's $8 / $40 list; cache reads $0.12, writes $6.00
Fable 5.2 Benchmarks
Anthropic: FrontierSWE v2 69.9% (Fable 5.1: 56.3%), Terminal-Bench-Science 74.6% (52.6%), Remote Labor Index 41.3% (17.9%)
Fable 5.2 vs Opus 5.5
Ahead on FrontierCode 2.0 (49.1% vs 45.5%) and Humanity's Last Exam with tools (74.2% vs 67.7%); about half the cost per task at Opus 5.5's best coding score
Faster, Fewer Refusals
Anthropic: outputs up to 30% faster than Fable 5.1 and about 55% fewer safety-classifier blocks
1M Context, 128K Output
Text and image input; same request surface as Fable 5.1, so switching is a one-line model change
Thinking Always On
Steer depth with effort (low to max, default high); forced tool use runs as auto plus an instruction
About Claude Fable 5.2
Claude Fable 5.2 is a Large Language Model API provided by Anthropic. Claude Fable 5.2 is Anthropic's most capable widely released model, released on 9 November 2026 as the successor to Fable 5.1, and on APIMODELS it costs $4.80 per 1M input tokens and $24.00 per 1M output — 40% below Anthropic's $8 / $40 list — with prompt-cache reads at $0.12 and cache writes at $6.00 for both the 5-minute and the 1-hour TTL (Anthropic charges $0.20, $10 and $16). Anthropic cut the list price by 20% from Fable 5.1's $10 / $50 and says outputs are up to 30% faster. On Anthropic's launch table it leads both Fable 5.1 and Opus 5.5 on agentic coding — FrontierCode 2.0 49.1% at xhigh effort (Fable 5.1 32.4%, Opus 5.5 45.5%) and FrontierSWE v2 69.9% (56.3%, 62.3%) — and on long-horizon and knowledge work: Terminal-Bench-Science 0.1 74.6% (52.6%, 58.7%), Remote Labor Index 41.3% (17.9%, 21.3%), Humanity's Last Exam with tools 74.2% (65.6%, 67.7%) and Chartography without tools 73.4% (44.8%, 64.4%). It is not first everywhere: Gemini 4 Carbon scores higher on AutomationBench (53.9% against 51.4%). Anthropic also reports that at high effort Fable 5.2 matches Opus 5.5's best FrontierCode score for roughly half the cost per task; both models are 40% below list here, so that ratio holds on APIMODELS too. Specifications: a 1M-token context window, 128K max output, and text and image input. Thinking is always on — control its depth with output_config.effort (low to max, default high). temperature, top_p, top_k and budget_tokens are removed on this model upstream, and APIMODELS drops them instead of forwarding them. Anthropic rejects forced tool use (tool_choice any or a named tool) on Fable 5.2; APIMODELS sends it as tool_choice auto plus an instruction to call that tool, which is also how structured outputs run here. Assistant prefill is not supported. Anthropic's safeguards decline most high-risk cybersecurity work, such as penetration testing and exploit generation, and those requests come back with stop_reason refusal. Call it through /v1/messages (Claude Code, the Anthropic SDK, Cursor) or the OpenAI-compatible /v1/chat/completions with one key; failed requests are not charged. Through APIMODELS platform, you can access this model via a unified API with transparent pay-as-you-go pricing. Current pricing: Input: $4.80, Output: $24.00 per 1M tokens.
Key Features
- 40% Below Official -- $4.80 / $24.00 per 1M tokens against Anthropic's $8 / $40 list; cache reads $0.12, writes $6.00
- Fable 5.2 Benchmarks -- Anthropic: FrontierSWE v2 69.9% (Fable 5.1: 56.3%), Terminal-Bench-Science 74.6% (52.6%), Remote Labor Index 41.3% (17.9%)
- Fable 5.2 vs Opus 5.5 -- Ahead on FrontierCode 2.0 (49.1% vs 45.5%) and Humanity's Last Exam with tools (74.2% vs 67.7%); about half the cost per task at Opus 5.5's best coding score
- Faster, Fewer Refusals -- Anthropic: outputs up to 30% faster than Fable 5.1 and about 55% fewer safety-classifier blocks
- 1M Context, 128K Output -- Text and image input; same request surface as Fable 5.1, so switching is a one-line model change
- Thinking Always On -- Steer depth with effort (low to max, default high); forced tool use runs as auto plus an instruction
Use Cases
Chatbot & Customer Support
Build intelligent conversational systems to automatically answer user queries and improve service efficiency.
Content Generation
Automatically write articles, emails, ad copy, and other text content to boost productivity.
Code Assistant
Assist with code writing, debugging, and code review to accelerate software development.
Data Analysis
Understand and analyze unstructured data, extract key insights, and generate summary reports.
Why APIMODELS
- Unified API -- One API key to access all models, no need to register on multiple platforms
- Transparent Pricing -- Pay-as-you-go with per-call pricing published on every model page, ideal for indie developers and startups
- Instant Access -- Start using immediately after signup, supports Stripe and Alipay payments
- Full Documentation -- Detailed API docs with code examples in cURL, Python, and Node.js
Frequently Asked Questions
How much does Claude Fable 5.2 cost?
Claude Fable 5.2 is available through APIMODELS at: Input: $4.80, Output: $24.00 per 1M tokens. Billing is pay-as-you-go — you only pay for what you generate.
How to use Claude Fable 5.2 API?
Sign up at APIMODELS, get your API key, and call our unified API endpoint. We provide detailed API documentation with code examples in cURL, Python, and Node.js.
What is the difference between APIMODELS and the official Anthropic API?
APIMODELS offers the same Claude Fable 5.2 model through our aggregation platform. We provide a unified API interface so you do not need separate accounts for each provider - one API key to access all models.
How much does Claude Fable 5.2 cost on APIMODELS?
$4.80 per 1M input tokens and $24.00 per 1M output tokens — 40% below Anthropic's $8 / $40 list. Prompt-cache reads are $0.12 per 1M (Anthropic: $0.20) and cache writes are $6.00 per 1M for both TTLs (Anthropic: $10 for 5 minutes, $16 for 1 hour), so passing ttl:"1h" costs nothing extra. The full 1M context bills at the same input rate, with no long-context surcharge. For reference, Fable 5.1 is $5.00 / $25.00 here, so the newer model is also the lower-priced one. Billing is per token on actual usage, and failed requests are not charged.
What are the Claude Fable 5.2 benchmarks?
From Anthropic's launch table (9 November 2026, safeguards on), with Fable 5.1 and Opus 5.5 in brackets: FrontierCode 2.0 49.1% at xhigh effort (32.4%, 45.5%); FrontierSWE v2 69.9% (56.3%, 62.3%); Terminal-Bench-Science 0.1 74.6% (52.6%, 58.7%); AutomationBench 51.4% (31.4%, 42.5%); Remote Labor Index 41.3% at high effort (17.9%, 21.3%); Humanity's Last Exam with tools 74.2% (65.6%, 67.7%); Chartography without tools 73.4% (44.8%, 64.4%); OSWorld 2.1 offline set 89.3% (83.5%, 87.2%). Two caveats: Gemini 4 Carbon is ahead on AutomationBench at 53.9%, and the OSWorld figure is partial credit per step, not the share of tasks fully completed. These are Anthropic's published numbers, not our own runs.
Claude Fable 5.2 vs Opus 5.5 — which should I use?
Per token, Opus 5.5 is exactly half the price here ($2.40 / $12.00 against $4.80 / $24.00), but the bill depends on what a finished task costs. Anthropic reports that Fable 5.2 at high effort scores 45.3% on FrontierCode 2.0, level with Opus 5.5's best of 45.5%, for about half the cost per task; at medium effort it matches Opus 5.5's top WANDR search score for about 40% of the cost; and at low effort it matches Opus 5.5 on Humanity's Last Exam for under a tenth. Both are 40% below list here, so those ratios carry over. A practical split: long agent runs, hard coding and research go to Fable 5.2, judged on cost per finished task; short, high-volume calls stay on Opus 5.5. For requests Fable 5.2's safeguards decline (mostly high-risk security work), Anthropic's own fallback is Opus 4.8, which is also on APIMODELS. They share one key and one endpoint, so switching is a change to the model field.
Should I move from Fable 5.1 to Fable 5.2?
For most workloads, yes. On APIMODELS it costs less ($4.80 / $24.00 against $5.00 / $25.00). Anthropic says outputs are up to 30% faster, that at medium effort it finishes hard tasks with fewer tokens than Fable 5.1, and it scores higher on every benchmark where the launch table has both models; safety-classifier blocks are down by about 55%. The request surface matches Fable 5.1 — thinking always on, no sampling parameters, no prefill — so switching is a one-line change to the model field. The one cost to plan for: prompt caches are kept per model, so the first request after switching writes the cache again.
How do I call Fable 5.2, and does Claude Code work?
Yes. On the native Anthropic path, set ANTHROPIC_BASE_URL to https://api.apimodels.app, set the model to claude-fable-5-2 and use your apimodels key (see /docs/claude-code); the Anthropic SDK and Cursor's custom Anthropic model setting connect the same way, with native tool use, streaming and prompt caching. With the OpenAI SDK, call POST https://api.apimodels.app/v1/chat/completions with model claude-fable-5-2 — reasoning_effort maps to Anthropic's effort, and tools and image input are translated for you. Single requests on hard tasks can run for several minutes, so stream long generations (stream: true) and give your client a generous timeout.
What are the limitations of Fable 5.2 on APIMODELS?
Worth knowing before you build on it. Thinking cannot be turned off — effort low is the lightest setting — and thinking {type: disabled} or budget_tokens are dropped. temperature, top_p and top_k are dropped too, so steer style through the prompt. Anthropic does not allow forced tool use on this model; we send it as tool_choice auto plus an instruction to call that tool, which picked the right tool in all 10 of our test requests but is not a decoding-level guarantee, so check that stop_reason is tool_use. Assistant prefill returns an error. Anthropic's Fast mode ($16 / $80) and Batch API are not offered here. Safeguards decline most high-risk cybersecurity requests (stop_reason refusal). And success is not 100%: if the primary upstream errors we switch to a second channel automatically, failed requests are not charged, and the model is too new for us to publish a success rate.
How does Claude Fable 5.2 compare to other Large Language Model models?
On APIMODELS, Claude Fable 5.2 runs alongside 60+ models on one API key and one balance, so choosing is about fit, not lock-in. It supports 40% Below Official, Long-Horizon Agents, 1M Context, Agentic Coding, and you can weigh it on price and capability against other Large Language Model models, then switch by changing a single model-name string — no new account or integration. Browse every Large Language Model option with live pricing at apimodels.app/models.
What can Claude Fable 5.2 do?
Claude Fable 5.2 supports: 40% Below Official, Long-Horizon Agents, 1M Context, Agentic Coding. See the APIMODELS docs for full parameters and call examples.
Can I access the Claude Fable 5.2 API from anywhere (incl. China)?
Yes. APIMODELS exposes Claude Fable 5.2 through a single unified API and one key — no separate provider accounts, and no need to handle each provider's regional network access yourself.
What payment methods are supported?
We support Stripe (Visa, Mastercard, and other international cards) and Alipay. Credits are available instantly after payment.
Related models
Related reading
How to get access, regional availability, and how this model compares with its alternatives.