Seed

41.4 Better than usual experience score ▲ 0.3 Confidence: Medium

68 opinions tracked this week · 7-day rolling window

30-day experience score

20304050602026-06-23 · Score jumped 21 pts vs prior 7-day average2026-06-24 · Score jumped 19 pts vs prior 7-day average2026-06-25 · Score jumped 14 pts vs prior 7-day average2026-06-26 · Score jumped 12 pts vs prior 7-day average2026-06-30 · Score dropped 14 pts vs prior 7-day average2026-07-01 · Score dropped 10 pts vs prior 7-day average06-1106-1606-2106-2607-0107-0607-10

Version breakdown

Only comments that explicitly name a version contribute to its score.

Doubao mentions only n=2 Confidence: Insufficient

66 generic family mentions are included in the family score, not any version score.

By experience dimension

General text 48 n=36
Coding not enough data n=0
Reasoning 32 n=19
Image / vision 55 n=5
Roleplay / creative not enough data n=4
Safety & refusals not enough data n=3
Speed & latency not enough data n=1
Local deploy not enough data n=0

Top positive drivers

  • competitor flight General text ×5
  • strong reasoning Reasoning ×4
  • emotional companionship Roleplay / creative ×3
  • problem solved General text ×3
  • creative surprise General text ×2

Top negative drivers

  • hallucination Reasoning ×10
  • logic error Reasoning ×3
  • safety overreach Safety & refusals ×3
  • competitor flight General text ×3
  • hallucination General text ×3

What users are saying

🔴 “速度:Hy3快,豆包无论什么模式现在都偏慢,长一点的搜索任务能感觉到拖沓。... 豆包那边经常出现'搜之前还算自然,搜完变成书面体简报',割裂感重。”

via zhihu · speed_latency · zh

🟢 “独角兽那个图宁愿选豆包。gpt也不过是模板花哨些罢了。”

via zhihu · image_vision · zh

🔴 “aistudio用半个小时我能tm得脑淤血,听不懂话且听懂了也做/讲不对”

via zhihu · text_general · zh

🟢 “豆包的实时语音大模型上线也有一年多了,随时打断、听懂你说话中间的停顿到底是没说完还是真说完了、嘈杂环境下不被背景音干扰,这些基本都已经是常规操作。”

via zhihu · text_general · zh

🟢 “目前豆包给我解过好几个卦都应验了。特别神奇。比如有次工作盘点...下班以后起卦问豆包和deepseek,东西丢没丢,在哪里放的。他俩都告诉我没丢,在低处,阴暗潮湿的地方,靠近金属架子,大概在西南方位。第二天上班按照提示果然找到了。”

via zhihu · reasoning · zh

How's your AI experience today?

One tap, no sign-up. Pick a model and tell us.