Pelican Bicycle ArchiveAN AI FIELD GUIDE
Community record · 2026-10-21

An SVG cyclist and a long coding session

Still preview

About this work

The author explicitly identifies Gemini 4 Argon and an SVG coding workflow. Two reported steps used 61,057 and 39,339 tokens. Quota and cost statements describe this account, not general pricing. The ajs6888 repost is not another output. Model attribution is source-reported and unverified. Full recording and source are retained; unpublished code is not mirrored and no onsite Play demo is offered.

Date evidence, prompt & provenance

Date precision / evidence: day / source-publication. Source publication: 2026-10-21. Generation date is not independently authenticated.

Recorded date
2026-10-21
Date precision
Day
Date basis
source-publication
Model (source label)
Gemini 4 Argon
Model attribution
Source-reported, not independently authenticated
Creator / publisher
码农暖爸 (@Delroy715)
Format
Animation
Prompt category
Source-documented test
Original prompt
一个「鹈鹕骑车」的任务,直接消耗 Google AI Pro 周额度的 3% 😂 最近看到不少朋友吐槽 Google AI Pro 给的 Gemini 4 Argon 额度太少,我干脆自己买了一个台区 Pro,折合约 33 元/月,实测一下到底能用多少。 这次直接让它深度思考,并生成完整的 SVG 代码,实测数据如下: 步骤 1:深度思考 + 完整 SVG 代码生成 消耗 61,057 tokens 步骤 2:写入完成 + 特性总结回复 消耗 39,339 tokens 📊 总消耗:100,396 tokens 💰 总成本(优惠期前): $0.47184 × 6.68 ≈ ¥3.15 特别注意: 1、只有周额度,没有 5 小时限制。 2、Colab 云端算力有限,目前 Pro 下最高只能开到 A100 80G。 3、96G 的 G4 实测两次都没开成功,暂时不确定是额度、资源还是其他限制导致的。 一个看起来不算复杂的小任务,就消耗了约 3% 的周额度。难怪不少朋友觉得 Pro 给的 Argon 量不够用。 后面再继续测测,这每月 33 元到底能撑多久。
Last updated
2026-10-22
Markdown record