Trending storyDeveloping
Claude Sonnet 4.5 benchmark scores and pricing
1 report1 sourceupdated 2 hours ago
What happened
AI digest
Epoch AI's evaluation page on October 3 shows Anthropic's closed-weight Claude Sonnet 4.5 at an ECI of 147, currently 66th of 253 tracked models; the page also records it as 4th of 166 models at release. On specific benchmarks it scores 71% on SWE-bench Verified and 82% on GPQA Diamond. Pricing is $3 per million input tokens and $15 per million output tokens.
Written by AI from the coverage · updated 2 hours ago
LatestOct 3 23:14
Epoch AI's page shows Sonnet 4.5 at an ECI of 147, now ranked 66th.Coverage
Follow the reports to see the story from different sides.
Oct 3
- AI HOT · ModelsEpoch AI 评测页面：Claude Sonnet 4.5 基准成绩与定价
Anthropic 的闭源模型 Claude Sonnet 4.5 在 Epoch AI 评测中取得 ECI 147 分，在 253 个追踪模型中排第 66 位，发布时在 166 个模型中排第 4。其定价为输入 $3 / 1M tokens、输出 $15 / 1M tokens，SWE-bench Verified 得分 71%，GPQA Diamond 得分 82%。
Heat over time
Not enough continuous observations to draw a trend yet.