Claude Fable 5
Flagshipby Anthropic
Claude Fable 5 is Anthropic's new frontier tier, announced June 9, 2026 and positioned above Claude Opus 4.8, which was the default model at the time. Claude Opus 5 has since landed at $5/$25, half the Fable 5 rate, and outscores it on most published rows. The API id is claude-fable-5, priced at $10 per million input tokens and $50 per million output with no long-context surcharge. It ships a 1 million token context window by default with up to 128K output tokens, takes text and vision input, and runs adaptive thinking always on (it cannot be disabled) with effort levels spanning low, medium, high, xhigh, and max. Anthropic's launch table reports 80.3 on SWE-bench Pro, 85.0 on OSWorld-Verified, 29.3 on FrontierCode Diamond, and a GDPval-AA ELO of 1932; treat all of these as vendor-reported until independent runs land. One operational note: always-on safety classifiers can reroute flagged requests to Opus 4.8, billed at Opus rates for the rerouted portion. Available on the Claude API, Amazon Bedrock, Google Vertex, and Microsoft Foundry.
Input Price
$10.00
per 1M tokens
Output Price
$50.00
per 1M tokens
Context Window
1M
tokens
Released
2026-06
API access
Capabilities
Key Strengths
- ✓1M token context window with 128K max output
- ✓Frontier tier above Opus 4.8
- ✓Adaptive thinking always on with five effort levels (low to max)
- ✓Vendor-reported 80.3 SWE-bench Pro and 85.0 OSWorld-Verified
- ✓Day-one availability on Claude API, Bedrock, Vertex, and Foundry
Best For
- ▸Hardest long-horizon agentic work
- ▸Whole-repository refactors
- ▸Multi-document research synthesis
- ▸Computer-use automation
Benchmark Scores
| Benchmark | Score | Description |
|---|---|---|
| SWE-bench | 95.0 | Real-world software engineering tasks from GitHub issues (SWE-bench Verified) |
| OSWorld 2.0 | 66.1 | Computer use across real desktop applications and multi-step GUI tasks |
| BrowseComp | 87.4 | Agentic web search and browsing over hard-to-find facts |
| FrontierCode v1.1 | 53.5 | Agentic coding on frontier software engineering tasks (Main split) |
| Humanity's Last Exam (tools) | 63.9 | Multidisciplinary expert-level reasoning with tool access |
Scores sourced from public benchmark datasets. See full benchmark leaderboard for all models.
Pricing Details
Input tokens
$10.00
per 1M tokens
Output tokens
$50.00
per 1M tokens
Estimated cost per 1K requests
$35.00
~1K input + ~500 output tokens avg
Prices are subject to change. Check the official documentation for current pricing. See the cost calculator for detailed estimates.
Related Models
Claude Opus 5
FlagshipAnthropic
$5.00 in / $25.00 out
Claude Opus 4.8
FlagshipAnthropic
$5.00 in / $25.00 out
Claude Opus 4.7
Mid-tierAnthropic
$15.00 in / $75.00 out
Claude Opus 4.6
FlagshipAnthropic
$15.00 in / $75.00 out
Claude Sonnet 5
Mid-tierAnthropic
$2.00 in / $10.00 out
Claude Sonnet 4.6
Mid-tierAnthropic
$3.00 in / $15.00 out
Claude Haiku 4.5
BudgetAnthropic
$0.80 in / $4.00 out
GPT-5.6 Sol
FlagshipOpenAI
$5.00 in / $30.00 out
GPT-5.5
FlagshipOpenAI
$5.00 in / $30.00 out
GPT-4o
FlagshipOpenAI
$2.50 in / $10.00 out
o1
FlagshipOpenAI
$15.00 in / $60.00 out