Google Deepmind 发布 Gemini 4 碳在Argon获得全面收益仅两周后
关键点
- Gemini 4 Argon后两周，Google DeepMind 发布 Gemini 4 Carbon，它的下一个前沿模型。 Carbon 超过人工分析智能指数v5 有59分，从Argon的50分上。
- Google 称 Carbon为多日自主开发工作流程所建。 它在人工分析的编码Agent指数(59分对Argon的50分)中快速地缩小了差距，但仍然跟随了Claude Opus 5.5 (62分).
- 由于其先进的网络安全能力，Carbon正在经历美国政府的自愿预测试过程，一开始只能供选定的安全专家使用。 定价和Argon一样
Google Deepmind在Argon刚刚两周后就将Gemini 4 Carbon运出，而新型号在人工分析智能指数上占据了首位。
碳在最近更新的人工分析智能指数5.0上得分59分，而前作"Argon踪迹"以50分。 Google除了在代理任务和编程方面普遍受益外，还突出了跨越整个网络防御生命周期的能力，从发现和验证弱点到补丁。
为多天编码会话所建
Google Deepmind 称它为需要持续代理工作流程的"许多更难，长视线的工程问题"设计了Carbon. 内部团队对模型进行了多天的直通运行，其间它独自通过死端来工作，并在大型代码库和低等软件Google 声称中处理了"根"等复杂多步骤的任务。
Google还报告在科学推理，多模式理解，以及基于用户反馈的全网开发方面的改进。 网络开发中质量更高的产出和文物将"碳"推高了"WebDev Arena"上89个埃洛分，以1,769分位列第四。
在Gray Swan的间接提示词注入基准测试上，Carbon将攻击成功率由以Argon为单位的0.7%推到0.0 %，阻止了按照Google进行的每次被测试的操纵尝试。 这并不意味着提示词注入是被解开的，但测试看起来是饱和的。
Google数字的独立评价
根据"人工分析","碳"在其最高推理级("高")领先"智能指数v5"，这主要是因为在复杂的推理，知识检索，和科学任务中得分最高，在指数的9个个别评价中占据了第一位。
在Antigravity上的代理编码中，Carbon仍然以相较于Argon从53起跳出到59起，以3分的分数跟随了Opus 5.5的Claude Code. 在关于AA-Briefcase的代理知识工作方面，碳(高)比Argon(高，1,483埃洛)获得126分，达到1,609埃洛，受展示质量(+293埃洛到1,598)和分析质量(+120埃洛到1,671)的驱动。 它仍然落后于Kimi K3.1 Pro(1,678 Elo),Muse Wonder(最大为1,659 Elo)和Grok 4.7 (x高为1,638 Elo).
与阿尔贡相比，答案精度从48%上升至58%，而幻觉率从4%上升至11%. 这更高，但仍是排名第二的。 Claude Opus 5.5 (最大)和GPT-6 Astra (最大)各占50%.
Vals AI的评价平台将Carbon的准确度排在第二位，为70.36%，落后于Claude Sonnet 5.5(70.85%)，领先于Claude Opus 5.5(70.18%)和Argon(69.24%)，成本均较低。
在Vals.ai上，Carbon还同时领先金融指数和法律研究基准测试。 在法律研究中，Vals AI记录了大约45%被引用的来源，更频繁的工具呼叫，以及比Argon更长的响应。
但"金融与税收"(Financial and Tax Agent 基准测试)中的任务需要与"阿尔贡"(Argon)相比的两倍多的时间，"法律研究"(Legal Research)中，"碳"(Carbon)平均每项任务需要超过50分钟。 该模型生成的推理符比Argon多出2.3倍，输出符比Argon多出64%，虽然这还不到Sonnet 5.5和Opus 5.5的输出符数的一半。
在Arena.ai上，Gemini 4 Carbon (High)以1,528分的成绩取了Text Arena领先的成绩，在Argon上行(1,520分). 在Agent Arena，该模型在超过21,900个真实会话中位居第二，领先于GPT-6 Astra和Fable-5.1，净提高16.2%，可引导性分数最高(+18.6%). 其1.8美元的任务成本中位数低于第一位Claude Opus 5.5 (Max)的72%.
在Code Arena的WebDev分类中，Carbon总共达到1,768分(比第11位的Argon领先89分)，仅落后于Claude Sonnet 5.5分(x High, 1,770分). 这使得它处于帕雷托曲线上，跨越了所有三个学科。
没有护栏的碳铅安全基准测试
在几次安全测试中，Google在没有安全护栏的情况下运行了该型号。 在SEC-bench Pro中，它测量现实世界软件中的自主脆弱性发现，碳得分为88.2%，并领导所有测试模型。 在测试一个系统能否将发现的弱点转化为工作开发的探索基因上，该模型达到了54.8%。 Carbon在CWE-Bench上还拥有广泛的第一领先的写作修饰。
Google还引用了无法独立核实的内部结果。 在内部的基准测试上，据报道Carbon生产的安全代码比Argon多出一倍多，在Google的克罗姆团队的测试中，它写安全补丁的表现比一个无名竞争者的前沿模型要好十多倍。
安保专家先进入
Google正在逐步推出该型号并参与美国政府的放行前测试程序，引用了Carbon的先进安全能力。 部分网络安全专家从本周开始首先接入。 通过Google Deepmind's Fairwind Program，之后访问范围会扩大到更多信任的伙伴。
公司计划在收集更多回馈并调整了安全护栏后，通过产品接口向开发商，企业和消费者开放。
Gemini 4 Conform 发布以与Argon相同的入门价格：每百万个输入令牌为2美元，每百万个输出令牌为10美元，其中缓存输入令牌得到95%的平价优惠。
人工分析情报指数中每个任务碳成本为7.76美元，比Claude Opus 5.5(最大，10.04美元)少了约23%，比GPT-6 Astra(最大，6.25美元)多出约1.2倍。 它平均每项任务有123,000个输出令牌，比Anthropic型号(173,000个令牌)少了29%，而获得几乎完全位于Argon的出价点(7.68美元).
人工分析的计算依据是每百万个输入令牌4美元和每百万个产出令牌20美元(缓存输入令牌外加0.20美元)的常规列表价格，而忽略了目前2026年底的50%入门折扣。
AI News Without the Hype – 由人类治愈
订阅THE DECODER免费阅读，每周AI通讯，我们独家的"AI雷达"前沿报告每年6次，全面存档访问，访问我们的评论部分。
现在订阅
本文由 AI宝库自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Google Deepmind releases Gemini 4 Carbon just two weeks after Argon with gains across the board
Key Points
- Just two weeks after Gemini 4 Argon, Google DeepMind released Gemini 4 Carbon, its next frontier model. Carbon tops the Artificial Analysis Intelligence Index v5 with 59 points, up from Argon's 50.
- Google says Carbon is built for multi-day, autonomous development workflows. It's closing the gap fast in Artificial Analysis's Coding Agent Index (59 points vs. Argon's 50) but still trails Claude Opus 5.5 (62 points).
- Because of its advanced cybersecurity capabilities, Carbon is going through the U.S. government's voluntary pre-testing process and will only be available to selected security experts at first. Pricing stays the same as Argon.
Google Deepmind shipped Gemini 4 Carbon just two weeks after Argon, and the new model takes the top spot on the Artificial Analysis Intelligence Index.
Carbon scores 59 points on the recently updated version 5.0 of the Artificial Analysis Intelligence Index, while predecessor Argon trails at 50. Beyond general gains in agentic tasks and programming, Google highlights capabilities spanning the full cyber defense lifecycle, from detecting and validating vulnerabilities to patching them.
Built for multi-day coding sessions
Google Deepmind says it designed Carbon for "much harder, long-horizon engineering problems" that demand persistent agentic workflows. Internal teams ran the model for multiple days straight, during which it worked through dead ends on its own and handled complex, multi-step tasks like root-cause analysis and subtle bug fixes in large codebases and low-level software, Google claims.
Google also reports improvements in scientific reasoning, multimodal understanding, and full-stack web development based on user feedback. Higher-quality outputs and artifacts in web development pushed Carbon up 89 Elo points on the WebDev Arena, where it sits in fourth place at 1,769.
On Gray Swan's indirect prompt injection benchmark, Carbon pushed the attack success rate from 0.7 percent with Argon down to 0.0 percent, blocking every tested manipulation attempt according to Google. That doesn't mean prompt injection is solved, but the test appears to be saturated.
Independent evaluations back up Google's numbers
According to Artificial Analysis, Carbon at its highest reasoning tier ("high") leads the Intelligence Index v5 mainly because of top scores in complex reasoning, knowledge retrieval, and scientific tasks, taking first place in four of the index's nine individual evaluations.
In agentic coding on Antigravity, Carbon still trails Claude Code with Opus 5.5 by three points, despite jumping from 53 to 59 compared to Argon. In agentic knowledge work on AA-Briefcase, Carbon (high) gains 126 points over Argon (high, 1,483 Elo) to reach 1,609 Elo, driven by presentation quality (+293 Elo to 1,598) and analytical quality (+120 Elo to 1,671). It still falls behind Kimi K3.1 Pro (1,678 Elo), Muse Wonder (max, 1,659 Elo), and Grok 4.7 (xhigh, 1,638 Elo).
Answer accuracy rises from 48 to 58 percent compared to Argon, while the hallucination rate climbs from four to 11 percent. That's higher but still the second-lowest in the ranking. Claude Opus 5.5 (max) and GPT-6 Astra (max) each sit at 50 percent.
The evaluation platform Vals AI puts Carbon in second place at 70.36 percent accuracy, behind Claude Sonnet 5.5 (70.85 percent) and ahead of Claude Opus 5.5 (70.18 percent) and Argon in fourth (69.24 percent), all at lower cost.
On Vals.ai, Carbon also leads both the Finance Index and the Legal Research benchmark. In legal research, Vals AI recorded roughly 45 percent more cited sources, more frequent tool calls, and longer responses than Argon.
But tasks in the Finance and Tax Agent benchmark take more than twice as long as with Argon, and in Legal Research, Carbon averages more than 50 minutes per task. The model generates 2.3 times as many reasoning tokens and 64 percent more output tokens than Argon, though that's still less than half the output token count of Sonnet 5.5 and Opus 5.5.
On Arena.ai, Gemini 4 Carbon (High) takes the Text Arena lead with 1,528 points, edging past Argon (1,520). In the Agent Arena, the model places second across more than 21,900 real sessions, ahead of GPT-6 Astra and Fable-5.1, with a net improvement of 16.2 percent and the highest steerability score (+18.6 percent). Its median task cost of $1.80 is 72 percent below first-place Claude Opus 5.5 (Max).
In the WebDev category of the Code Arena, Carbon reaches 1,768 points (89 ahead of eleventh-place Argon) for fourth overall, just two behind Claude Sonnet 5.5 (xHigh, 1,770). That places it on the Pareto curve across all three disciplines.
Carbon leads security benchmarks without guardrails
For several security tests, Google ran the model without safety guardrails. On SEC-bench Pro, which measures autonomous vulnerability discovery in real-world software, Carbon scores 88.2 percent and leads all tested models. On ExploitGym, which tests whether a system can turn discovered vulnerabilities into working exploits, the model reaches 54.8 percent. Carbon also holds a wide first-place lead in writing fixes on CWE-Bench.
Google also cites internal results that can't be independently verified. On in-house benchmarks, Carbon reportedly produces secure code more than twice as often as Argon, and in tests by Google's Chrome team, it performed more than ten times better at writing security patches than the frontier model of an unnamed competitor.
Security experts get access first
Google is rolling out the model gradually and participating in the U.S. government's pre-release testing process, citing Carbon's advanced security capabilities. Selected cybersecurity experts get access first starting this week. Through Google Deepmind's Fairwind Program, access will expand to more trusted partners after that.
The company plans to open the model to developers, businesses, and consumers through its product interfaces after collecting more feedback and adjusting the safety guardrails.
Gemini 4 Carbon launches at the same introductory price as Argon: $2 per million input tokens and $10 per million output tokens, with cached input tokens getting a 95 percent discount off the regular input price.
Carbon costs $7.76 per task in the Artificial Analysis Intelligence Index, about 23 percent less than Claude Opus 5.5 (max, $10.04) and roughly 1.2 times more than GPT-6 Astra (max, $6.25). It averages 123,000 output tokens per task, 29 percent fewer than the Anthropic model (173,000 tokens), and lands almost exactly at Argon's price point ($7.68).
Artificial Analysis bases its calculations on the regular list price of $4 per million input tokens and $20 per million output tokens (plus $0.20 for cached input tokens), ignoring the current 50 percent introductory discount that runs through the end of 2026.
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
Subscribe now
Auto-fetched and translated by AI Gem Box. All rights belong to the original author.