Breaking

Anthropic ships Claude Fable 5.1 with stronger coding and 75% cheaper cache reads

September 2, 2026 at 12:47 EDT

Foundation Models
Software Dev & Coding
AI Agents

Anthropic on September 1, 2026 released Claude Fable 5.1, a generally available Mythos-class model for coding and knowledge work, keeping Fable 5’s $10 per million input and $50 per million output rates while cutting cache-read prices 75% to $0.25 per million tokens. The company says that change makes typical token-billed workloads about 25% cheaper than Fable 5 and highly agentic ones up to about 45% cheaper. A restricted twin, Claude Mythos 5.1, shares the same underlying model with more permissive cybersecurity and life-sciences safeguards.

September 1, 2026 · Anthropic

Fable 5.1: same sticker, cheaper agents

Claude Fable 5.1 keeps Fable 5’s $10 / $50 per million token rates and cuts cache reads 75% to $0.25, with stronger long-horizon coding scores.

Cache read price (per million tokens)

$1.00

Fable 5

$0.25

Fable 5.1

Four stacked blocks versus one: cache hits cost a quarter of the Fable 5 rate.

~25%

Typical workload saving vs Fable 5

~45%

Highly agentic estimated saving

55.8%

Terminal-Bench 4.0 (Fable 5: 42.0%)

Where testers see gains

Longer unattended coding, prototypes in days, and cache-heavy agent loops that finally undercut prior Fable bills when hits dominate.

Caveats

List input and output did not fall. Without cache reuse, the sticker is unchanged; some max-effort task costs rose with more output tokens.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.