Artificial Analysis Index 推出 v5 更新，Claude Opus 5.5 居首
评估榜单 Artificial Analysis 更新到 v5，Claude Opus 5.5 超过 GPT-6 系列，选模型前可以看看这版新排名。
Artificial Analysis Index 发布 v5 版本，此前版本因使用过旧基准导致结果偏差而受批评。新版排名中 Claude Opus 5.5 领先 GPT-6 Astra 和 GPT-6.1 Sol，Google 的 Gemini for Argon 排在 Anthropic 和 OpenAI 之后。Kimi K 3.1 Pro 列开源权重模型第一，尽管其权重尚未放出。基准方面用 HLE Diamond 替换 Humanity's Last Exam，Artificial Analysis 还将 AA OmniScience 更新到 1.1，并新推出用于对比长期能力变化的 Progress Index v1。
Artificial Analysis Index gets a much-needed version 5 update. The previous version had been criticised for using very old benchmarks that gave us very odd results.
The upgrade confirms Claude Opus 5.5 as the leader ahead of GPT-6 Astra and GPT-6.1 Sol. Interestingly Gemini for Argon puts Google as the third lab behind Anthropic and OpenAI.
Kimi K 3.1 Pro is the top open-weight model even though the weights have not been released. Open-weight models generally fall further behind closed models.
HLE Diamond replaces Humanity's Last Exam in the index. Opus 5.5 leads again. They also updated their AA OmniScience to version 1.1.
And they're launching the Artificial Analysis Progress Index, version 1, which compares underlying capability over time.
