Anthropic's Opus 5 out-muscles Fable 5 at half the price, benchmark slapdown ensues 🥊

Anthropic released Claude Opus 5 on Monday, positioning the lower-tier model as a replacement for Claude Fable 5 in everyday paid use while undercutting it on price. Opus 5 outperformed Fable 5 across most benchmarks reviewed and, in several cases, beat OpenAI's GPT-5.6 Sol by wide margins. Anthropic's lineup runs four tiers: Haiku (fast and cheap), Sonnet (mid-range), Opus (heavy workhorse), and the Mythos class introduced this spring, which includes Claude Fable 5 for the public and Claude Mythos 5, a less-restricted version reserved through Project Glasswing for vetted cybersecurity researchers and critical infrastructure operators.

Fable 5 launched June 9, was pulled globally three days later after the U.S. government issued an emergency export control order citing a jailbreak vulnerability, and returned June 30 before shifting immediately to a credits-only model outside standard plans. Opus 5 now occupies the slot Fable 5 could not hold.

On Frontier-Bench v0.1, which tests end-to-end software engineering tasks scored as a percentage passed, Opus 5 hit 43.3%, compared with 33.7% for Fable 5 and 34.4% for GPT-5.6 Sol. The widest margin came on ARC-AGI-3, a problem-solving test built around novel puzzles scored as a percentage solved: Opus 5 reached 30.2% versus 7.8% for GPT-5.6 Sol, with Fable 5 not tested. On GDPval-AA v2, a knowledge-work benchmark scored via Elo ratings, Opus 5 reached 1,861 against Fable 5's 1,747 and GPT-5.6 Sol's 1,736.

Lovable, a developer platform with millions of users, reported gains beyond raw scores. "It isn't just better on our hardest agentic coding tasks, up 22% over Opus 4.7, it's steadier, with far less variance run to run," Fabian Hedin said in a statement shared by Anthropic. Zapier's AutomationBench, which scores whether a model can carry a full business workflow without human help, returned a clean result: the model "took a raw account-health workbook and ran a full churn-prevention sequence end to end: flagging at-risk accounts, alerting the right owner, and summarizing for retention ops. Previous models didn't pass; Opus 5 hit 100%."

Anthropic is also pitching Opus 5 as a research upgrade, with Ultima Genomics, a DNA sequencing company, stating the model "behaves more like a careful scientist than any model we've run" in reaching for the right statistical tests, cross-checking results, and ruling out confounders.