nafyi.

PUBLIC VIEWPOINTS

Simon Willison

Source author ·

1 interviews · 3 viewpoints · 3 topics

Content updated:

Explore connections ↗

Translations are for reading; original excerpts remain the evidence.

Viewpoints by topic

Attributed viewpoints, ordered by interview date. A snapshot of these conversations, not a definitive statement of someone’s beliefs.

AI model performance and cost

View this topic
AI model performance and cost

Reported Sonnet 5.5 speed and cost improvements

Willison quotes Anthropic’s claim that Sonnet 5.5 "runs 30%+ faster, and costs up to 30% less for most work". He writes that it is priced the same as Sonnet 5, "but appears to beat it on every benchmark, and should be cheaper to run as well".

Supporting evidence

Original excerpt

Claude Sonnet 5.5 . New Sonnet model from Anthropic today. They say it "runs 30%+ faster, and costs up to 30% less for most work" - it's priced the same as Sonnet 5 but appears to beat it on every benchmark, and should be cheaper to run as well.

Simon Willison·

Claude Sonnet 5.5

Additional context is not included in this excerpt. Read the original conversation.

Open original · Paragraph 2 (opens in a new tab)Original text source (opens in a new tab)

Text location from the supplied original; confirm wording and attribution in context.

Simon Willison
Share insight

AI model reliability and token limits

View this topic
AI model reliability and token limits

Token exhaustion in a pelican-SVG test

In Willison’s pelican-SVG test, Sonnet 5.5 at “max” thinking effort used 128,000 tokens, costing $1.28, before running out of tokens without producing an SVG.

Supporting evidence

Original excerpt

the "max" thinking effort pelican thought for 128,000 tokens (at a cost of $1.28) before running out of tokens and failing to produce an SVG.

Simon Willison·

Claude Sonnet 5.5
Context

Here are some pelicans riding bicycles . Sonnet 5.5 suffered from the same bug as Opus 5.5 :

Open original · Paragraph 3 (opens in a new tab)Original text source (opens in a new tab)

Text location from the supplied original; confirm wording and attribution in context.

Simon Willison
Share insight

AI model capability comparison

View this topic
AI model capability comparison

Sonnet 5.5 near-parity with Opus 5.5 on coding tasks

Willison says Sonnet 5.5 appears almost as good as Opus 5.5 on some coding tasks, including viral 3D animation tricks.

Supporting evidence

Original excerpt

Sonnet 5.5 appears to be almost as good as Opus 5.5 on some coding tasks, including various viral 3D animation tricks .

Simon Willison·

Claude Sonnet 5.5

Additional context is not included in this excerpt. Read the original conversation.

Open original · Paragraph 5 (opens in a new tab)Original text source (opens in a new tab)

Text location from the supplied original; confirm wording and attribution in context.

Simon Willison
Share insight

Statements by interview date1

Statements are ordered by the original interview publication date; differences in wording do not establish a change of position.