Unverified50% confidenceBenchmarkExact time
用 Opus 5 在 Medium 设置下配合 Ponytail 运行基准测试，输出 Token 明显减少，成本实际便宜约 22%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/22/2026
First Seen
Valid until: 10/20/2026
Sources
Related Claims
UnverifiedThrough a series of strategy adjustments, the usage cost of Opus 5 can be reduced by up to 80%, while still outperforming Opus 4.8 on some benchmarks.73% similarUnverifiedIn the Artificial Analysis Intelligence Index comprehensive benchmark, GPT-5.6's performance is approximately 0.9 points lower than Claude Opus 4.5, but at nearly half the cost.72% similarUnverifiedPonytail's benchmark tests show a reduction in token consumption of 22%.70% similarUnverifiedPonytail's official benchmark uses weaker models, but testing with Opus and stronger models yields results that are even cheaper, faster, and better—meaning the stronger the model, the greater the benefit.68% similarUnverified在编程智能体测试中，Sol 输出的 Token 少了一半以上，成本约为 Faber-5 的 1/368% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/593992API
curl https://kongchang.com/api/v1/knowledge/claims/593992MCP
get_claim(id=593992)