Anthropic releases Claude Opus 5.5 at $4 and $20 per million tokens, 20% below Opus 5, and it takes first place on the Artificial Analysis Intelligence Index

Anthropic released Claude Opus 5.5 on 22 September and calls it the first model in a new Claude 5.5 family. Its announcement gives the price as $4 per million input tokens and $20 per million output tokens, “20% less than Opus 5”, which charged $5 and $25. Cache reads, which Anthropic says make up most of the cost of agentic and coding work, fall to $0.20 per million tokens.

Anthropic also gives a 40% figure, which is a different measure: it says the model “costs 40% less to run than Opus 5”, at default settings on typical workloads in its own tests. The list price is 20% lower.

Anthropic claims Opus 5.5 performs at the level of Claude Fable 5.1 on most work. Its own table puts it at 66.4% on Terminal-Bench 4.0, against 52.3% for Opus 5. The model is on Anthropic's platforms and the big clouds, including Amazon Web Services, Google Cloud and Microsoft Azure, and under its safeguards most cybersecurity tasks are routed to Opus 4.8 instead.

The independent benchmarker Artificial Analysis published its own write-up the same day, headed “Claude Opus 5.5 takes the top spot on the Artificial Analysis Intelligence Index”. At maximum effort the model scores 58 on that index. About 90 minutes later OpenAI launched GPT-6 Sol and GPT-6 Luna, TechCrunch reported.

• Confirmed Anthropic lists Opus 5.5 at $4 and $20 per million input and output tokens, “20% less than Opus 5”. Anthropic
• Claimed Anthropic says the model “costs 40% less to run than Opus 5” at default settings. Anthropic
• Confirmed Artificial Analysis says Opus 5.5 scores 58 at max effort and takes the top spot on its Intelligence Index. Artificial Analysis
• Reported OpenAI launched GPT-6 Sol and Luna about 90 minutes after Anthropic's release, TechCrunch reported. TechCrunch