Fable
An Anthropic model derived from the Mythos tier, with capabilities reduced relative to the original Mythos specification.
Core Facts
Claude is Anthropic's AI large language model.
In the deep audit benchmark, Sonnet 5.5's cost is approximately half that of Opus, its score is slightly higher than Opus, and it takes about 5 minutes (roughly half the time Opus takes).
Accounting for real usage scenarios, the actual cost advantage of Opus 5 over Fable is approximately 20% to 25%, with Fable costing about $2.75 per task and Opus 5 about $2.03.
Timeline (last 90 days)
In Claude Code, the Haiku model supports a context of about 256,000 tokens, while Opus and Fable reach about 1 million tokens.
Official benchmarks show Kimi K3.1 Pro scoring 65.7 on Terminal Bench, second only to Opus 5.5's 66.4; 1846 ELO on GDP Val versus Opus 5.5's 1866; and 79.6 on OS World, on par with Opus and Fable.
The pattern of a simple prompt plus free exploration may be more effective in complex codebases than fine-grained tuning.
One developer spent $48.62 using Anthropic's Fable feature to solve a critical bug that had plagued the team for months.
Claude.ai uses a subscription plus pay-as-you-go model (for features such as Fable), with clear billing details.
Price transparency may be more important than the absolute price, because with predictable costs the team was willing to pay $48.62.
Fable allows users to describe a problem in natural language and lets the model autonomously explore, locate and modify code in the codebase, using an agentic workflow.
After the fix, the result is still not very smooth, but it is much better than before.
On Vibe Code Bench, Argon scored 91.9 while Opus and Fable scored 90.3, so Argon's advantage is slight and covers only one type of coding task.
Claude Code v2.1.301 fixes the issue where model aliases (such as fable, best) error out when the organization has not enabled the latest Fable model; now it falls back to the previous Fable model and retries once.
40 more timeline events
All Facts (20)
Claude is Anthropic's AI large language model.
70%VerifiedIn the deep audit benchmark, Sonnet 5.5's cost is approximately half that of Opus, its score is slightly higher than Opus, and it takes about 5 minutes (roughly half the time Opus takes).
65%VerifiedAccounting for real usage scenarios, the actual cost advantage of Opus 5 over Fable is approximately 20% to 25%, with Fable costing about $2.75 per task and Opus 5 about $2.03.
65%VerifiedOpenAI's naming system shifted during the GPT-4 era from decimal version numbers to letter suffixes (such as 4-turbo and 4o) to identify architectural optimizations.
65%Partially VerifiedSonnet 5.5 is on par with Opus 5.5 in offensive and defensive cybersecurity capabilities, making it the first model in the Sonnet series to adopt equivalent cybersecurity protection mechanisms.
65%UnverifiedIn Claude Code, the Haiku model supports a context of about 256,000 tokens, while Opus and Fable reach about 1 million tokens.
50%UnverifiedOfficial benchmarks show Kimi K3.1 Pro scoring 65.7 on Terminal Bench, second only to Opus 5.5's 66.4; 1846 ELO on GDP Val versus Opus 5.5's 1866; and 79.6 on OS World, on par with Opus and Fable.
50%UnverifiedClaude.ai uses a subscription plus pay-as-you-go model (for features such as Fable), with clear billing details.
50%UnverifiedThe pattern of a simple prompt plus free exploration may be more effective in complex codebases than fine-grained tuning.
50%UnverifiedAfter the fix, the result is still not very smooth, but it is much better than before.
50%UnverifiedPrice transparency may be more important than the absolute price, because with predictable costs the team was willing to pay $48.62.
50%UnverifiedFable allows users to describe a problem in natural language and lets the model autonomously explore, locate and modify code in the codebase, using an agentic workflow.
50%UnverifiedOn Vibe Code Bench, Argon scored 91.9 while Opus and Fable scored 90.3, so Argon's advantage is slight and covers only one type of coding task.
50%UnverifiedAnthropic's average daily commit volume grew from 2,500 to 30,000, a jump that corresponds to the timing of internal deployments of Fable- and Mythos-class models.
50%UnverifiedThis developer's token usage by model is distributed as: Sonnet 82%, Opus 14%, Fable 3.6% (almost entirely concentrated within one day).
50%UnverifiedArgon is positioned by Google as a super-large model on the same level as Astra and Fable.
50%UnverifiedDeveloper Pavel Hurin tested Argon with a bug hunter task, and the result was comparable to Muse Spark 1.3, clearly lagging behind Astra, Fable and Opus.
50%UnverifiedFable level refers to a tier of model standard previously considered within Anthropic to be at a higher capability level.
50%UnverifiedAnthropic's 'Teamer' mode is coordinated by a strong model, with cheap models executing batch work; in one test, the total cost of Fable leading 25 Sonnet workers was about half that of using Fable alone, with time reduced from more than 15 hours to less than 2.5 hours, and the quality score 10 to 12 points lower.
50%UnverifiedFable is an AI-driven creativity and content generation tool that focuses on helping users quickly generate structured narrative content, including story creation, brand copywriting, and interactive experience design.
50%