Anthropic just dropped Claude Fable 5.1, and the upgrade reads like a developer wish list: it beats its own predecessor on coding benchmarks while getting cheaper on exactly the workloads that inflate API bills. The Fable 5.1 model arrived on September 1, three months after the original Fable 5, and the pricing move is the part that matters most for Filipino developers and startups.
What Claude Fable 5.1 is
In Anthropic’s lineup, Fable is Claude’s most capable tier, with Haiku, Sonnet, and Opus sitting below it. The official announcement frames Fable 5.1 as setting “a new standard on coding, knowledge work, and long-running problem-solving tasks.” It defaults to High effort in Claude Code and Medium in Claude Cowork and on claude.ai.
It is available today on Amazon Web Services, Google Cloud, and Microsoft Azure, with the model exposed as claude-fable-5-1 on the Claude API.
Up to 45% cheaper on agentic work
The headline number is the bill. Anthropic cut cache read pricing by 75%, down to $0.25 per million tokens, which is where long agentic runs spend most of their money. The company estimates Fable 5.1 costs about 25% less than Fable 5 for typical workloads, and up to roughly 45% less for context-heavy, tool-heavy work.
| Scenario | Fable 5.1 vs Fable 5 |
|---|---|
| Typical workloads | About 25% cheaper |
| Highly agentic work | Up to about 45% cheaper |
| Cache reads (API) | $0.25 per million tokens, 75% less |
A jump in coding scores
On Anthropic’s published benchmarks, Fable 5.1 roughly doubles Fable 5 on agentic scientific research and pulls ahead on agentic coding and business workflows.
| Benchmark | Fable 5.1 | Fable 5 |
|---|---|---|
| Terminal-Bench-Science 0.1 | 52.6% | 24.7% |
| Terminal-Bench 4.0 (agentic coding) | 55.8% | 42.0% |
| CursorBench 3.2.0 | 73.4% | 70.5% |
| AutomationBench | 31.4% | 17.1% |
Anthropic says the model “avoids shortcuts” and fixes root causes. In one cited test, investment firm Millennium reported that Fable 5.1 found the cause of a rare crash its engineers could not explain after four to five years of trying.
Mythos 5.1 and the safety changes
Alongside Fable 5.1, Anthropic released Claude Mythos 5.1, the same underlying model with stricter safeguards, limited to trusted access programs for cybersecurity and life sciences. For enterprises, a new Enterprise Frontier Safeguards system promises zero data retention with customer-controlled cloud infrastructure, rolling out this fall. Anthropic also says its cyber safeguards now block 60% fewer false positives.
What it means in the Philippines
Claude is a daily driver for many PH developers, agencies, and startups, and cheaper agentic pricing matters because context-heavy coding is exactly where monthly bills balloon. For local teams, this lands as an upgrade to the model that recently overtook Fable 5 on value, and it comes weeks after Anthropic adjusted Claude Code’s usage limits. Full benchmark details and availability are in the official announcement.
