Wire flash
Claude Sonnet 5.5 (High) scores 1699, ranks 4th in Code Arena WebDev
Editorial responsibility
- No named human review is recorded for this page.
- Source reporting is collected, normalized, translated or condensed automatically when needed.
- Automatically published source-backed update
According to a post on X by Arena (@arena), the Claude Sonnet 5.5 (High) model achieved a score of 1699 in the Code Arena: WebDev benchmark, placing it 4th overall. The mixed price is approximately $8 per million tokens, which is 80% cheaper than the 2nd and 3rd ranked models. Compared to the previous Sonnet 5 (High) model's score of 1540, this represents a 159-point improvement. The model also saw significant gains in subcategories: Reference-Based Design, Simulations, and Gaming, all rising from around 30th place to 4th place. The source provides benchmark rankings, scores, and price comparisons, allowing readers to evaluate the model's cost-performance position.
Source report
Arena has released benchmark results showing Claude Sonnet 5.5 (High) achieving 4th place in the Code Arena: WebDev category with a score of 1,699 points.
Pricing and Cost Comparison
- Mixed price: Approximately $8 per million tokens
- 80% cheaper than the 2nd and 3rd place models
Performance Improvement
Compared to its predecessor, Sonnet 5 (High) (score: 1,540 points), the new model shows a 159-point improvement.
Category Rankings
In the following subcategories, Claude Sonnet 5.5 (High) jumped from 30th+ place to 4th place:
- Reference-Based Design
- Simulations
- Gaming
Key Takeaway
The original report provides actual benchmark rankings, scores, and price comparisons, allowing readers to evaluate the model's position in terms of both cost and performance.
Source
aihotNeutral / independent