11-19 更新 实时
首页/ AI 大模型/ 正文

Gemini 4 Carbon 登顶 MathArena,成本是 GPT-6.1 Sol 的 15 倍

AI 大模型 来源:The Decoder 2026-11-15 00:48 0 阅读 机器翻译

Google 的 Gemini 4 Carbon 成为一段时间以来首个登顶 MathArena 数学基准测试的非 OpenAI 模型。它把 GPT-6.1 Sol 从榜首位置挤了下来,而 Claude Fable 5.2 位列第二,这使其有可能成为数学任务上能力最强的公开可用模型之一。OpenAI 的模型据称最近解出了 Navier-Stokes 方程这类高难度方程,但目前尚未公开可用。

Gemini 4 Carbon 最常见的失效模式是超出时间限制。该模型也几乎从不对任务放弃,这一特点既可以解读为优点,也可以解读为缺点。有一次,它启动了一个脚本,运行了超过 11 个小时都没有检查其进度。不过,这种额外的性能表现是有代价的。在测试所用的推理设置下,Carbon 和 Fable 5.2 的成本分别约为 GPT-6.1 Sol 的 15 倍和 40 倍。

不带炒作的 AI 新闻——由人工筛选

译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。

查看英文原文(Gemini 4 Carbon takes the MathArena crown but costs 15 times more than GPT-6.1 Sol)

Google's Gemini 4 Carbon has become the first non-OpenAI model in a while to top the MathArena math benchmark. It knocked GPT-6.1 Sol from the top spot, while Claude Fable 5.2 came in second, making it potentially one of the most capable publicly available models for math tasks. OpenAI's models, which recently allegedly solved demanding equations like the Navier-Stokes equations, aren't publicly available yet.

Gemini 4 Carbon's most common failure mode is exceeding the time limit. The model also almost never gives up on a task, a trait that can be read as either positive or negative. In one case, it launched a script that ran for more than eleven hours without checking on its progress. That extra performance comes at a price, though. With the reasoning settings used in testing, Carbon and Fable 5.2 cost roughly 15 and 40 times as much as GPT-6.1 Sol, respectively.

AI News Without the Hype – Curated by Humans

原文为英文,正文由自动抓取并保留原文。

本文由采集器自动同步

原文发布于 The Decoder(Gemini 4 Carbon takes the MathArena crown but costs 15 times more than GPT-6.1 Sol),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。

阅读原文 ↗