Anthropic has unveiled Claude Opus 5, a new GenAI model that the company says delivers performance on par with Claude Fable 5 while cutting token costs by around half.
The launch comes as organisations balance AI performance and costs, with Anthropic positioning Opus 5 as a lower-cost route to frontier AI.
Opus 5 marks a significant step up from Opus 4.8
According to Anthropic, Opus 5 is "substantially stronger" than Opus 4.8, with key improvements in agentic coding, computer use and long-horizon knowledge work.
The company says the model delivers "greatly improved performance" while maintaining the same pricing as Opus 4.8.
Anthropic also claims Opus 5 matches the performance of Claude Fable 5 and Mythos 5, placing it among the company's top-performing AI models.
Lower token prices combine with top-tier intelligence
Claude Opus 5 is priced at $5 per MTok for input tokens and $10 per MTok for output tokens, with a reported maximum task cost of $2.03.
The model tops the Artificial Analysis Intelligence Index with a score of 61.
By comparison, Claude Fable 5 has a reported task cost of $2.75 and an index score of 60, while OpenAI's GPT-5.6 Sol has a reported task cost of $1.04 and an index score of 59.
Meanwhile, Moonshot AI's Kimi K3 has an average task completion cost of $0.95.
GPT-5.6 Sol remains the fastest model in the comparison, generating 273 output tokens per second - around 220 more per second than Opus 5.
Anthropic cuts prompt overhead to speed up responses
Anthropic says it has removed 80% of system prompts from Opus 5, simplifying instructions and helping deliver results more quickly.
The company noted that responses can become longer as a result and recommends using explicit response-length instructions within prompts.
Cost reductions could help organisations scale AI deployments
Unlike traditional software licensing, AI models create variable operational costs that increase with usage. As organisations expand AI across multiple workflows, token consumption can become a significant factor in overall spend.
By reducing token pricing while maintaining high performance, Opus 5 could help organisations scale AI initiatives more cost-effectively.
Anthropic also believes the model could prove valuable for scientific research, with internal testing indicating improvements over Opus 4.8 in supporting research-focused workflows.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
Destination AI
Destination AI from TD SYNNEX is our comprehensive end-to-end programme designed to help you and your customers get ahead of the AI curve.
