Claude Sonnet 5 underperforms on CursorBench, lacks savings
A recent post points out that Claude Sonnet 5 not only performs terribly on CursorBench but also lacks its expected price advantage. The model's "Max" tier is priced only $0.72 cheaper than the flagship Claude Opus 4.8 Max, defeating the original purpose of the Sonnet lineup, which is designed to be a faster and significantly more affordable alternative.
Anthropic may have fundamentally misunderstood its user base by pricing a mid-tier model so close to its flagship.
- –Historically, the Sonnet series won developers over with a compelling price-to-performance ratio.
- –If Sonnet 5 Max is barely cheaper than Opus 4.8 Max, developers have almost no financial incentive to use it, especially given the reported poor benchmark performance.
- –CursorBench is a crucial benchmark for coding tasks, meaning poor results here could severely hamper adoption among software engineers.
DISCOVERED
3h ago
2026-06-30
PUBLISHED
5h ago
2026-06-30
RELEVANCE
AUTHOR
bridgemindai