Gemini 4 Neon
A model in Google's Gemini 4 series whose specific positioning remains unclear; it may be an upgraded version or part of a new series.
Core Facts
Timeline (last 90 days)
如果 Gemini 4 Neon 具备与 Opus 5.5 竞争的实力,开发者的模型选型将更依赖具体任务表现,而非品牌惯性。
「Gemini 4 Neon」的命名带有推测成分,尚无官方确认。
Hacker News 上出现了一条标题为「Gemini 4 Neon looks like an Opus 5.5 competitor」的讨论帖,获得 18 分和 1 条评论。
目前关于 Gemini 4 Neon 与 Opus 5.5 对比的所有判断均建立在一条简短帖子的标题之上,缺乏可验证的基准数据或产品细节。
Before the official documentation is completed, using Gemini 4 Neon for critical business systems still carries risk.
The official blog post for Gemini 4 Neon is categorized under the "Innovation & AI / Models & Research" section.
Gemini 4 Neon offers four reasoning levels: low, medium, high, and xhigh.
The official documentation explicitly lists coding, agent workflows, and multi-step tasks as key focus areas for Gemini 4 Neon.
The model ID for Gemini 4 Neon in AI Gateway is google/gemini-4-neon.
Gemini 4 Neon supports tool calling, structured outputs, and grounding capabilities based on Google Search and Google Maps.
17 more timeline events
All Facts (20)
Google published a new member of the Gemini series, Gemini 4 Argon, on its official blog.
90%Partially VerifiedThe identifiers Gemini 4 Neon and Gemini 4 Oxygen appear in Gemini Web backend logs, alongside Gemini 4 Argon and Carbon.
75%UnverifiedThe model ID for Gemini 4 Neon in AI Gateway is google/gemini-4-neon.
60%Unverified目前关于 Gemini 4 Neon 与 Opus 5.5 对比的所有判断均建立在一条简短帖子的标题之上,缺乏可验证的基准数据或产品细节。
50%Unverified「Gemini 4 Neon」的命名带有推测成分,尚无官方确认。
50%UnverifiedBefore the official documentation is completed, using Gemini 4 Neon for critical business systems still carries risk.
50%UnverifiedThe official blog post for Gemini 4 Neon is categorized under the "Innovation & AI / Models & Research" section.
50%UnverifiedGemini 4 Neon supports tool calling, structured outputs, and grounding capabilities based on Google Search and Google Maps.
50%UnverifiedGemini 4 Neon offers four reasoning levels: low, medium, high, and xhigh.
50%UnverifiedThe official documentation explicitly lists coding, agent workflows, and multi-step tasks as key focus areas for Gemini 4 Neon.
50%UnverifiedGemini 4 Neon supports streaming calls via the AI SDK's streamText.
50%UnverifiedOn SWE-sweep, Gemini 4 Neon scores 12.8% under the mini-SWE harness and 39.0% in Google Antigravity.
50%UnverifiedGemini 4 Neon sits on the Pareto frontier on Vals Index and Agent Arena, with cost per task about 50% to 80% lower than top competitors.
50%UnverifiedGemini 4 Neon scored 61 on the Artificial Analysis Intelligence Index at the xhigh thinking level, ranking third overall.
50%UnverifiedGemini 4 Neon ranked second on the Agent Arena Index with a score of 1336, an evaluation covering more than 4,200 real agent sessions.
50%UnverifiedGemini 4 Neon ranked first on both Legal Research Bench and Tax Agent Bench.
50%UnverifiedIn Surge AI's evaluation, Gemini 4 Neon ranked first in multiple categories including finance, accounting, spreadsheet modeling, and complex constraint following.
50%UnverifiedGemini 4 Neon scored 56.3% on Zapier's AutomationBench and $16,843 on Andon Labs' Vending-Bench 2 (a long-horizon business simulation), taking first place overall on both.
50%UnverifiedGemini 4 Neon scores 70.5% on Terminal-Bench Science, 84.0% on RiemannBench, and 71.3% on HLE-Diamond, leading in all.
50%UnverifiedGemini 4 Neon ranked second on the Vals Index with a score of 70.9%.
50%