1 min read
Add as a preferred source on Google

Opus 5.5 Cuts Token Prices 20% and Scores 64.8% on Terminal-Bench

Anthropic’s Sept. 22 release has a 1M-token context window, up from 200K for Opus 4.5, released Nov. 24, 2025.

A terminal keyboard rests beside a thick software migration printout / TokenPost.ai
A terminal keyboard rests beside a thick software migration printout / TokenPost.ai

Anthropic’s Opus 5.5 combines lower token prices with a larger context window and a 64.8% score on a command-line benchmark, giving developers new cost and performance figures to weigh when choosing an AI model.

Released Sept. 22, 2026, Opus 5.5 costs $4 per million input tokens and $20 per million output tokens. Those rates are 20% below Opus 4.5’s prices of $5 and $25, respectively. Opus 4.5 was released Nov. 24, 2025.

The context window grew from 200K tokens for Opus 4.5 to 1M for Opus 5.5. Cache reads for the newer model are priced up to 60% lower.

On Terminal-Bench 4.0, which tests command-line task performance, Opus 5.5 scored 64.8%. Opus 4.8 scored 23.6% on the same benchmark. The comparison covers two releases and does not establish performance across other tasks.

Anthropic says Opus 5.5 performs at the level of Fable 5.1 and costs about 40% less to run than Opus 5. The company also describes the model as matching or exceeding higher-tier Fable models on several benchmarks.

Early testers described completing complex coding migrations in hours rather than weeks, including one project involving 680,000 lines of code. Anthropic says Opus 5.5 posted its highest scores to date on its automated behavioral alignment audit and was evaluated externally before release.

As of mid-November 2026, Anthropic’s API offered Opus versions from 4.5 through 5.5.

Simon Yoon

Reporter

Simon Yoon reports on blockchain technology for TokenPost. Send corrections or tips to info@tokenpost.com.

Loading…