Anthropic released Claude Opus 5 on July 24, 2026 — its fourth Claude 5-generation model release in under two months, a pace that reflects how AI competition has shifted from blockbuster capability jumps toward rapid, iterative improvements in cost and efficiency.
What Anthropic Is Actually Claiming
The core pitch is straightforward: Opus 5 delivers intelligence approaching Anthropic’s top-tier Claude Fable 5 model across many tasks, while costing half as much. Pricing stays unchanged from its predecessor at $5 per million input tokens and $25 per million output tokens — the same as Opus 4.8 — meaning you’re getting a meaningfully more capable model without paying more for it.
On specific benchmarks covering coding and knowledge work, including Frontier-Bench and GDPval-AA, Anthropic says Opus 5 is now state-of-the-art among its models — though it notably still trails the specialized Mythos 5 model on cybersecurity-specific tasks.
Where It Fits in Anthropic’s Lineup
Opus 5 becomes the new default model for Claude Max subscribers and is now the strongest model available on Claude Pro. It’s positioned as Anthropic’s everyday, daily-use model for enterprises, developers, and knowledge workers — distinct from Fable 5, which remains the top-of-the-line option for the most demanding tasks at a higher price point ($10/$50 per million tokens).
The New “Effort Dial”
One practical new feature: users can now adjust how much computing power Opus 5 devotes to a given task through an effort control. At lower effort settings, Anthropic says the model preserves much of its performance while using fewer tokens — directly translating to lower cost for tasks that don’t need maximum reasoning depth. A related new option lets users switch which underlying model handles a task mid-conversation, giving more direct control over ongoing AI costs.
Real-World Performance Claims
Several companies building on Claude shared specific results following early access:
- Zapier reported Opus 5 topped its AutomationBench leaderboard, completing a full account churn-prevention workflow end-to-end without using more tokens than prior Claude models — a task previous models couldn’t fully complete.
- Cognition (makers of the Devin coding agent) reported Opus 5 approaches Fable-level performance at half the cost on coding benchmarks, with particular strength in debugging and root-cause analysis.
- Fundamental Research Lab reported roughly 9 percentage points higher accuracy on hard financial-modeling tasks, using about a third fewer tool calls and 60% less time than comparable models.
Availability
Opus 5 is available immediately across Anthropic’s platforms — the API (as claude-opus-5), Claude.ai, Claude Code, and Claude Cowork — as well as through Amazon Bedrock and Claude Platform on AWS, where it includes zero data retention by default for enterprise governance requirements.
Why This Release Pattern Matters
Axios noted this launch reflects a broader shift in how AI labs are competing: rather than chasing headline-grabbing capability jumps every few months, the race has increasingly turned toward optimizing cost, speed, and reliability for models people actually use daily. With Claude reportedly holding around 40% of the enterprise large language model market by usage as of early 2026, this cost-efficiency focus appears aimed squarely at defending and growing that enterprise position against competitors.
Frequently Asked Questions
Is Claude Opus 5 more expensive than Opus 4.8?
No — pricing is identical at $5 per million input tokens and $25 per million output tokens, while performance has improved significantly.
How does Opus 5 compare to Claude Fable 5?
Opus 5 approaches Fable 5’s intelligence on many everyday tasks at roughly half the cost, though Fable 5 (and the specialized Mythos 5) remain ahead for the most demanding, specialized work.
Do I need to change my API integration to use Opus 5?
If you’re migrating from Opus 4.8, the API structure and Claude Code integration carry over largely unchanged — it’s primarily a model-ID swap to claude-opus-5.
