Anthropic Launches Claude Opus 5, Matching Fable 5 Performance at Half the Cost

Anthropic has launched Claude Opus 5, its latest flagship artificial intelligence model, claiming it delivers performance close to Claude Fable 5 while costing 50% less to use. The launch comes just weeks after the company introduced Claude Fable 5 and amid growing competition from low-cost AI models developed by Chinese companies such as Moonshot AI's Kimi K3 and GLM 5.2.

According to Anthropic, Claude Opus 5 offers significant improvements over Claude Opus 4.8, particularly in coding, reasoning and knowledge intensive tasks. The company also says the model matches or outperforms Claude Fable 5 across several benchmarks while maintaining the same pricing as Opus 4.8, making it a more cost-effective option for developers and enterprise users.

Claude Opus 5 Targets High Performance at Lower Cost

Anthropic said Claude Opus 5 is designed to deliver frontier-level AI performance while reducing inference costs. The company claims the model provides performance comparable to Claude Fable 5 in several real-world applications, including software development, coding, and knowledge work, while allowing users to optimise performance through adjustable effort settings that balance intelligence, speed and token usage.

Benchmark Performance

According to Anthropic, Claude Opus 5 outperformed Claude Fable 5 on 10 of 14 benchmark tests and delivered performance close to Fable 5 on the remaining evaluations. The company also said Opus 5 surpassed OpenAI's GPT-5.6 Sol in selected benchmarks and more than doubled the performance of Claude Opus 4.8 on the Frontier-Bench v0.1 benchmark while maintaining the same operating cost.

On CursorBench 3.2, Anthropic said Opus 5 finished within 0.5% of Claude Fable 5's score while operating at approximately half the cost.

Claude AI Model Comparison

Model	Input Cost (per 1M tokens)	Output Cost (per 1M tokens)	Key Positioning
Claude Opus 5	$5	$25	Near Fable-level performance at lower cost
Claude Fable 5	$10	$50	Anthropic's highest-performing frontier model
Claude Opus 4.8	$5	$25	Previous flagship model

Real-World Use Cases

Anthropic highlighted several early testing examples to demonstrate Opus 5's capabilities. According to the company, the model successfully wrote its own computer vision pipeline to reconstruct a machine component into a 3D FreeCAD model and helped an engineer at a trading firm build a market data feed for a new exchange during a single development session.

These demonstrations were presented as examples of the model's ability to perform complex engineering and software development tasks with minimal user intervention.

Cybersecurity Safeguards Remain in Place

Unlike Claude Fable 5 and Claude Mythos, Anthropic said Claude Opus 5 was not specifically trained for offensive cybersecurity capabilities. The company stated that it intentionally limited training in sensitive cyber domains following earlier concerns around advanced AI models.

According to Anthropic, Opus 5 can assist developers in identifying and fixing software vulnerabilities, but the model includes safeguards designed to block requests involving high-risk cyber activities and other dual-use applications.

Pricing Comparison

One of the biggest differentiators for Claude Opus 5 is its pricing. Anthropic has kept pricing unchanged from Claude Opus 4.8, while making it substantially cheaper than Claude Fable 5.

Claude Pricing Comparison

Model	Input Tokens	Output Tokens
Claude Opus 5	$5 / million	$25 / million
Claude Fable 5	$10 / million	$50 / million
Moonshot Kimi K3	$3 / million	$15 / million

Although Opus 5 is significantly cheaper than Claude Fable 5, Anthropic acknowledged that Chinese open-weight models such as Kimi K3 continue to offer lower pricing.

Availability

Anthropic confirmed that Claude Opus 5 is now the default model for Claude Max subscribers and is also available as the most capable model for Claude Pro users.

Why It Matters

The launch comes as AI companies increasingly compete on both performance and cost efficiency. With enterprises closely monitoring AI spending, lower inference costs have become an important competitive advantage. By offering performance close to Claude Fable 5 at roughly half the price, Anthropic is positioning Claude Opus 5 as a more affordable option for developers, businesses and enterprise customers seeking advanced AI capabilities.

What is Claude Opus 5?

Claude Opus 5 is Anthropic's latest AI model, designed to deliver performance close to Claude Fable 5 while costing significantly less to use.

How much does Claude Opus 5 cost?

Claude Opus 5 is priced at $5 per million input tokens and $25 per million output tokens, matching the pricing of Claude Opus 4.8.

Is Claude Opus 5 cheaper than Claude Fable 5?

Yes. Claude Opus 5 costs half as much as Claude Fable 5, whose pricing starts at $10 per million input tokens and $50 per million output tokens.

Does Claude Opus 5 support cybersecurity tasks?

Anthropic said Claude Opus 5 can assist with identifying and fixing software vulnerabilities but was not trained for advanced offensive cybersecurity capabilities and includes safeguards against high-risk requests.

Where is Claude Opus 5 available?

Claude Opus 5 is available as the default model for Claude Max subscribers and as the strongest available model for Claude Pro users.