🔥維護記錄

1 Like

话说今儿咋没每周总结了

1 Like

我也在想…合理懷疑AI偷懶睡著了。

2 Likes

論壇已接入 grok-3-fast-latest
直接 @ grok 即可。

1 Like

论坛私聊AI中报错信息显示"model": “gemini-2.0-pro-exp”
既不是2.5也不是flash-thinking……为什么?

而且显示是FreeTier,配额非常短缺,每天3点恢复完,不到晚上就被人用尽了,能否给这里也设置成轮询,或者自动换成配额比较多的flash-thinking?

报错信息如下:
Sorry, it looks like our system encountered an unexpected issue while trying to reply.

Error details

{
“error”: {
“code”: 429,
“message”: “You exceeded your current quota, please check your plan and billing details. For more information on this error, head to: https://ai.google.dev/gemini-api/docs/rate-limits.”,
“status”: “RESOURCE_EXHAUSTED”,
“details”: [
{
@type”: “type.googleapis.com/google.rpc.QuotaFailure”,
“violations”: [
{
“quotaMetric”: “generativelanguage.googleapis.com/generate_content_free_tier_requests”,
“quotaId”: “GenerateRequestsPerDayPerProjectPerModel-FreeTier”,
“quotaDimensions”: {
“location”: “global”,
“model”: “gemini-2.0-pro-exp”
},
“quotaValue”: “25”
}
]
},
{
@type”: “type.googleapis.com/google.rpc.Help”,
“links”: [
{
“description”: “Learn more about Gemini API quotas”,
“url”: “https://ai.google.dev/gemini-api/docs/rate-limits
}
]
},
{
@type”: “type.googleapis.com/google.rpc.RetryInfo”,
“retryDelay”: “26s”
}
]
}
}

2 Likes

已切換為flash-thinking

1 Like

已重構,應該短暫下線了幾分鐘,剛剛。

1 Like


这什么

1 Like

好像是redirect的问题:face_with_crossed_out_eyes:

1 Like

論壇模型已更新為
gemini-2.5-flash-preview-04-17

Model RPM TPM RPD
Gemini 2.5 Flash Preview 04-17 1,000 1,000,000 10,000
1 Like

| 高考默寫 | AI課文 || AI高考 | AI默寫 | AI論語 | AI紅樓 |

已全部更新為gemini-2.5-flash-preview-04-17

1 Like

新模型的数学能力很好,但是写作能力似乎有点一言难尽……

1 Like

新模型好像不会爆俄罗斯文了。

1 Like

会的,天尊,会的……

“以及学习 конкретной 诗的好处”

1 Like

寫入一個新的AI:

區別:

grok-3-fast-latest
grok-3-mini-latest
這兩個是模型,用了全名,未設置頭像。

模型區別:

  • grok-3-fast-latest :這是 Grok 的「旗艦」模型 (flagship)。它擅長處理企業級的複雜任務,比如數據提取、寫程式、文本摘要,並且擁有較深的專業領域知識(像金融、醫療、法律、科學)。但 Has reasoning: No (沒有推理能力)——這可能指的是它不提供像 grok-3-mini 那樣明確可追蹤的「思考過程」,而不是完全沒有邏輯能力。
  • grok-3-mini-latest :這是個「輕量級」模型 (lightweight)。圖片上說它會在回應前先「思考」(thinks before responding),並且 Has reasoning: Yes (有推理能力),適合基於邏輯和不需要深厚專業知識的任務。雖然是 Mini 版,但在某些基準測試(如 AIME)上得分反而比 Grok 3 更高,但在其他方面(如 GPQA, LCB)較低,各有擅長。它的成本也比 Grok 3 低很多。

Grok Grok
Grok-3-Mini Grok 3 Mini
這兩個是 AI 對話 bot,設置頭像,日常使用時直接 @ 了聊天即可。

2 Likes

已更新,啟用新輸入框。

Supported features

The following features are supported in the new composer:

  • Ability to add title and set categories and tags
  • Single-pane editor when in rich text editor mode (i.e. no preview pane)
  • Toggle to switch between Markdown-only and rich text editor modes
  • Basic formatting supported in Markdown, keyboard shortcuts, and toolbar icons
  • Emoji support
  • Auto-converting Markdown to rich text when pasting plain text (e.g., **bold**bold)
  • Auto-convert emoticons (e.g., :):slight_smile: )
  • Table builder UI
  • Poll builder UI
  • Footnotes
  • Support for mentions and hashtags
  • Inline HTML <tags>
  • Typographer replacements (e.g., :trade_mark: , → , ± )
  • Image uploads
  • File attachments
  • Topic and chat quotes
  • Hide Details
  • Blur Spoiler (with limitations, see Missing Features section)
  • Oneboxes
  • Horizontal rule autoconversion (i.e. typing --- converts to horizontal rule)
2 Likes

https://i.rdfzer.com/
加入暗夜模式,但你得自己猜點哪裡。

5 Likes

点一下北大附校徽

2 Likes


今天是生民以來300k yrs誒

1 Like

头像怎么微缩了()

1 Like