Google 发布 Gemini 4 Neon:Artificial Analysis 排名第 3,价格低 32%
Google 新模型 Gemini 4 Neon 能打:编码智能体榜第 1,4,200 个真实智能体会话里排第 2,价格还比竞品便宜一半。正在跑智能体或编码工作流的开发者可以算下成本账。
Google 推出 Gemini 4 Neon,在 Artificial Analysis Intelligence Index 上得分 61,排名全榜第 3,仅次于更大、更慢的模型。定价为每百万 token 输入 3 美元、输出 15 美元。默认高设置下,Neon 单任务成本比 Gemini 4 Carbon 低 32%,在 Vals Index 和 Agent Arena 上比顶级竞品前沿模型便宜约 50% 到 80%。
具体能力上,Neon 在 Artificial Analysis Coding Agent Index 以 58.6 分排名第 1;在 Google Antigravity 环境中运行时,SWE-sweep 成绩从 12.8% 提升到 39.0%。智能体任务方面,Neon 在 Zapier AutomationBench 以 56.3% 排名第 1,在 Vending-Bench 2 上取得 16,843 美元成绩排名第 1,并在超过 4,200 个真实智能体会话的 Agent Arena 中排名第 2。
对开发者而言,Neon 把前沿级智能体和编码性能压到明显更低的价格区间,与 Gemini 4 Carbon 及其他厂商旗舰模型相比成本优势直接体现在每任务花费上,适合高并发的智能体工作流和编码场景。
Google just launched Gemini 4 Neon, a new model ranking 3rd on Artificial Analysis.
> Scores 61 on the Artificial Analysis Intelligence Index, 3rd overall behind only larger, slower models, at $3/$15 per 1M input/output tokens.
> At its default high setting, Neon costs 32% less per task than Gemini 4 Carbon and roughly 50% to 80% less than top rival frontier models on the Vals Index and Agent Arena.
> In coding, Neon ranks 1st on the Artificial Analysis Coding Agent Index at 58.6. Its SWE-sweep score jumps from 12.8% to 39.0% when run inside Google Antigravity.
> Ranks 1st on Zapier’s AutomationBench at 56.3% and on Vending-Bench 2 with $16,843, and 2nd on Agent Arena across 4,200+ real agent sessions.
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Gemini 4 Neon, Google, coding assistants, reasoning models.
Source attribution: curated from rohanpaul_ai (@rohanpaul_ai).