Claude Sonnet 5.5 Just Dropped

Here is the catalog entry for the video:

Summary

Akinyemi Bajulaiye reviews the launch of Anthropic's Claude Sonnet 5.5 model, walking through the official release announcement, benchmark scores, and pricing details. He highlights the model's significant improvements in agentic coding over both Claude Sonnet 5 and Claude Opus 5.5, while noting its lower operating costs and increased speed.

What is shown

• [00:00] Official Anthropic announcement landing page for "Claude Sonnet 5.5" (dated September 28, 2026).
• [00:11] Benchmark comparison table detailing performance metrics across Claude Sonnet 5.5, Claude Sonnet 5, Claude Opus 5.5, and OpenAI's GPT-6 Sol.
• [01:08] Anthropic launch blog text outlining key feature updates, architectural context within the Claude 5.5 family, and alignment safeguards.
• [01:17] Official posts from Claude's X (formerly Twitter) account summarizing launch highlights, followed by community reaction posts.
• [01:28] Performance vs. cost curve graph on Terminal-Bench 4.0.
• [01:40] Pricing comparison table showing token costs for Sonnet 5.5 versus Opus 5.5 alongside sample code artifact demos.

Claims & numbers

• Benchmarks & Performance:
  • The presenter and displayed table claim Claude Sonnet 5.5 achieves 70.6% on Terminal-Bench 4.0 (agentic coding), beating Claude Opus 5.5 (66.4%), GPT-6 Sol (49.2%), and Claude Sonnet 5 (10.3%).
  • On CursorBench 4.0, Sonnet 5.5 scores 55.5% compared to Sonnet 5's 34.1%, Opus 5.5's 57.8%, and GPT-6 Sol's not available.
  • On FrontierCode-1.1 (Multi), Sonnet 5.5 scores 43.2%, beating Sonnet 5 (4.5%), Opus 5.5 (34.4%), and GPT-6 Sol (not available).
  • Knowledge work (AA Briefcase v1.7): Sonnet 5.5 scores 1831, Opus 5.5 scores 1822, and GPT-6 Sol scores 1483.
  • Multidisciplinary reasoning (Humanity's Last Exam): Sonnet 5.5 scores 64.5% without tools, compared to Opus 5.5 at 67.7%.
  • Computer use (OSWorld 2.1): Sonnet 5.5 reaches 60.1% partial, while Opus 5.5 reaches 61.8% partial.
• Speed & Efficiency:
  • Sonnet 5.5 runs 30%+ faster and costs up to 30% less for most work than Sonnet 5.
• Pricing:
  • Sonnet 5.5 pricing is $2 per million input tokens and $10 per million output tokens (cache reads $0.20, cache writes $2.50).
  • Opus 5.5 pricing is shown as $4 per million input tokens and $20 per million output tokens (cache reads $0.20, cache writes $5.00).
• Upcoming Events:
  • The presenter claims OpenAI DevDay is scheduled for tomorrow, and rumors indicate a new frontier Gemini model launch may be imminent.

Notable quotes

• [00:13] "It actually beats out Opus 5.5 on agentic coding."
• [00:23] "That's compared to Sonnet 5, which was 10% on the terminal bench, so this is a real step up from what we've seen."
• [01:43] "So Sonnet 5.5 is half the price of Opus 5.5: $2 for input tokens... $10 for output tokens."

Assessment

This is an independent creator commentary and reaction video covering an official release, not a live hands-on benchmark execution. The presenter analyzes Anthropic's published tables and marketing materials without independently executing the benchmarks or testing the model live on screen.

Described by gemini-3.8-flash on 2026-09-29 from the video's audio and frames.