Claude Opus 5: Near-Frontier Performance for Everyday Use
Anthropic released Claude Opus 5 on July 24, 2026, the latest flagship model in the Opus lineup. Priced at the same rate as Opus 4.8—$5 per million input tokens and $25 per million output tokens—it delivers substantially better performance across the board, approaching Fable 5's capabilities at half the cost.
Performance Highlights
The benchmark numbers are straightforward. On Frontier-Bench v0.1, Opus 5 surpasses all other models, more than doubling Opus 4.8's performance at a lower cost per task. On CursorBench 3.2 at max effort, it scores within 0.5% of Fable 5's peak while costing half as much per task. On ARC-AGI 3, which tests a model's ability to solve novel problems, Opus 5 scores three times higher than the next-best model.
On OSWorld 2.0, a computer use benchmark, Opus 5 outperforms every other model at any given cost, surpassing Fable 5's best result at just over a third of the cost. On Zapier's AutomationBench leaderboard, Opus 5 took the top spot without spending more tokens—it processed a raw account-health workbook and ran a complete churn-prevention workflow end to end.
Agent Capabilities
Opus 5 is noticeably better at verifying its own work and iterating until it succeeds. In one Frontier-Bench task, the model was given a drawing of a machine part and asked to rebuild it as a 3D FreeCAD model, without being given direct access to view the drawing. Opus 5 wrote its own computer vision pipeline to extract geometry from raw pixels, then reconstructed the full part. In another task, it found the root cause of a real bug in a popular open-source package manager and fixed an edge case that the community patch had missed.
Internal evaluations show Opus 5 up 22% over Opus 4.7 on the hardest agentic coding tasks, with significantly lower variance run to run.
Scientific Research
Opus 5 shows meaningful improvements over Opus 4.8 across all life sciences evaluations. The biggest gains are in organic chemistry—inferring molecular structures from spectroscopy data improved by 10.2 percentage points. Protein-related tasks, like predicting how sequence variations affect function, are up 7.7 percentage points. Early-access users describe Opus 5 as behaving "more like a careful scientist," reaching for appropriate statistical tests and cross-checking results by independent methods.
Safety and Availability
Opus 5 does not advance the frontier in risky dual-use capabilities—it remains behind Mythos 5 in both biology research and offensive cybersecurity. It's Anthropic's most aligned model to date, with better adherence to Claude's Constitution than Opus 4.8, Sonnet 5, or Fable 5, and the lowest rates of deceptive behavior.
Opus 5 is available today on all platforms: as claude-opus-5 on the API, as the default model on Claude Max, and as the strongest model on Claude Pro. Fast mode runs at approximately 2.5x default speed at double the base price.
