swyx 评 Gemini 4 Carbon 登顶 AA 智能指数：前沿实验室数量翻倍

Google 的 Gemini 4 Carbon 在 Artificial Analysis Intelligence Index v5 上成为新的第 1 名，在 Artificial Analysis Coding Agent Index 上则仅次于 Claude Opus 5.5 Gemini 4 Carbon 在 Intelligence Index v5 上得分 59，这得益于其在复杂推理、知识和科学任务上的领先表现。它在编程和知识工作方面也有大幅提升，尤其是与其他 Gemini 模型相比，但它并未在所有这些领域都处于领先地位。 与 Gemini 4 Argon 发布时一样，Gemini 4 Carbon 正在向部分用户推出，尚未公开可用。Google 提供 50% 的发布折扣，但未确认结束日期。我们所有的图表和数据均使用 Gemini 4 Carbon 和 Argon 的标准定价。 Gemini 4 Carbon 高推理模式的关键结果： ➤ Gemini 4 Carbon 在 Intelligence Index v5 上领先，在 9 项 Index 评估中的 4 项上获得第 1 名：HLE-Diamond、Terminal-Bench-Science、AA-Omniscience、AutomationBench-AA。Carbon 在其他方面也取得了大幅提升，但在编程和知识工作上并未领先 ➤ 在编程智能体上仅次于 Claude Opus 5.5：借助 Antigravity CLI，Gemini 4 Carbon（高）在 Artificial Analysis Coding Agent Index 上得分 59，仅落后于使用 Claude Opus 5.5（max，62）的 Claude Code 3 分。Carbon 在 Terminal-Bench 4.0（64%）和 FrontierCode v1.1（51%）上均排名第二，落后于 Claude Opus 5.5（max）的 56%。Google 在这一领域历来表现不佳，因此这一提升尤为突出 ➤ 智能体知识工作有所改善，但尚未领先：在 AA-Briefcase 上，Gemini 4 Carbon（高）得分为 1609 Elo，比 Gemini 4 Argon（高，1483）提升了 126 分，但仍落后于 Grok 4.7（xhigh，1638）、Muse Wonder（max，1659）和 Kimi K3.1 Pro（1678）等模型。最大提升来自 Presentation，上升 293 Elo 至 1598，同时 Analytical Quality 提升了 120 分至 1671 ➤ 在整体智能相当的情况下，每任务成本低于 Claude Opus 5.5：Gemini 4 Carbon 每个 Intelligence Index 任务的成本为 $7.76，比 Claude Opus 5.5（max，$10.04）低约 23%，但约为 GPT-6 Astra（max，$6.25）的 1.2 倍。它每个任务使用 123k 输出 token，比 Claude Opus 5.5（max，173k）少 29%。在相同的 $4/$20 标准定价下，其每任务成本与 Gemini 4 Argon（$7.68）相当，而在 Intelligence Index 上得分高出 9 分 ➤ AA-Omniscience v1.1 得分领先，且幻觉率低：Gemini 4 Carbon 在我们最近更新的 AA-Omniscience v1.1 上得分为 54，比 Claude Opus 5.5（max，48）高 6 分，比 Gemini 4 Argon（46）高 8 分。准确率（% 正确）较 Argon 提升 10 分至 58%，而其幻觉率从 4% 上升至 11%。这仍是所有模型中第二低的，仅次于 Argon，相比之下 Claude Opus 5.5（max）和 GPT-6 Astra（max）均为 50% 关键模型细节： ➤ 上下文窗口：1M tokens ➤ 多模态：文本、图像、视频和语音输入，文本输出 ➤ 定价：每 1M 输入/输出 tokens 为 $4/$20，与 Gemini 4 Argon 的标准定价相同，并享有 50% 折扣。缓存输入 tokens 享受 95% 折扣（每 1M $0.20） 祝贺 @GoogleDeepMind 团队！

查看原文

Google’s Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5, and second only to Claude Opus 5.5 on the Artificial Analysis Coding Agent Index Gemini 4 Carbon scores 59 on Intelligence Index v5, driven by leading performance in complex reasoning, knowledge, and scientific tasks. It also has large gains in coding and knowledge work, particularly when compared to other Gemini models, but it is not leading in all these areas. As with launch of Gemini 4 Argon, Gemini 4 Carbon is being rolled out to selected users and is not yet publicly available. Google is offering a 50% launch discount, with no confirmed end date. All our charts and figures use standard pricing for both Gemini 4 Carbon and Argon. Key results for Gemini 4 Carbon with high reasoning: ➤ Gemini 4 Carbon leads the Intelligence Index v5, with the #1 score in 4 of the 9 Index evaluations: HLE-Diamond, Terminal-Bench-Science, AA-Omniscience, AutomationBench-AA. Carbon makes substantial gains elsewhere but is not leading in coding and knowledge work ➤ Behind only Claude Opus 5.5 on coding agents: With Antigravity CLI, Gemini 4 Carbon (high) scores 59 only 3 points behind Claude Code with Claude Opus 5.5 (max, 62) on the Artificial Analysis Coding Agent Index. Carbon ranks second on both Terminal-Bench 4.0 (64%) and FrontierCode v1.1 (51%), behind Claude Opus 5.5 (max) at 56%. Google has historically underperformed in this area and so this improvement stands out ➤ Improved agentic knowledge work, but not yet leading: On AA-Briefcase, Gemini 4 Carbon (high) scores 1609 Elo, a 126-point improvement over Gemini 4 Argon (high, 1483), but still behind models including Grok 4.7 (xhigh, 1638), Muse Wonder (max, 1659) and Kimi K3.1 Pro (1678). The largest gain comes from Presentation, up 293 Elo to 1598, alongside a 120-point improvement in Analytical Quality to 1671 ➤ Lower Cost per Task than Claude Opus 5.5 for comparable overall intelligence: Gemini 4 Carbon costs $7.76 per Intelligence Index task, ~23% less than Claude Opus 5.5 (max, $10.04) but ~1.2x GPT-6 Astra (max, $6.25). It uses 123k output tokens per task, 29% fewer than Claude Opus 5.5 (max, 173k). At the same $4/$20 standard pricing, its Cost per Task is in line with Gemini 4 Argon ($7.68), while scoring 9 points higher on the Intelligence Index ➤ Leading AA-Omniscience v1.1 score, with low hallucination: Gemini 4 Carbon scores 54 on our recently updated AA-Omniscience v1.1, 6 points ahead of Claude Opus 5.5 (max, 48) and 8 points ahead of Gemini 4 Argon (46). Accuracy (% correct) rises 10 points over Argon to 58%, while its hallucination rate rises from 4% to 11%. This is still the second lowest of any model, behind only Argon, compared with 50% for both Claude Opus 5.5 (max) and GPT-6 Astra (max) Key model details: ➤ Context Window: 1M tokens ➤ Multimodality: Text, image, video, and speech input, with text output ➤ Pricing: $4/$20 per 1M input/output tokens, the same as Gemini 4 Argon’s standard pricing, with a 50% discount. Cached input tokens receive a 95% discount ($0.20 per 1M) Congratulations to the @GoogleDeepMind team!