Gemini
▲ 0.6
63.6Sharper than usualexperience score
313 voices this week Confidence: High
- n=135
- Hacker News
- n=113
- Zhihu · RedNote · V2EX
- n=65
Gemini 3.7 Flash is outperforming the family by 17.4 #1 in Speed & latency
Who's dumber today — and who's not
Every model off its own baseline right now.
Need an alternative? See what works best right now →
1,174 opinions tracked in the last 24h · 3 reported here directly Sources: Reddit · Hacker News · Zhihu · RedNote · V2EX
Updated 2026-08-16T10:26:58Z UTC
How's your AI experience today?
One tap, no sign-up. Pick a model and tell us.
Experience score, 0–100 · higher = more positive community experience · 7-day window
GPT
▲ 2.3
51.8As usualexperience score
1,881 voices this week Confidence: High
- n=1,525
- Hacker News
- n=126
- Zhihu · RedNote · V2EX
- n=230
GPT-5.6 Luna is outperforming the family by 11.7
Claude
–
42.2As usualexperience score
2,471 voices this week Confidence: High
- n=1,985
- Hacker News
- n=267
- Zhihu · RedNote · V2EX
- n=219
Claude Fable 5 is outperforming the family by 15.2 #1 in Reasoning
Grok
▲ 1.1
27.6Dumber than usualexperience score
633 voices this week Confidence: High
- n=491
- Hacker News
- n=72
- Zhihu · RedNote · V2EX
- n=70
Grok 4.6 is outperforming the family by 33.9
Best by task right now
GLM
▼ 0.8
69.9Sharper than usualexperience score
260 voices this week Confidence: High
- n=123
- Hacker News
- n=40
- Zhihu · RedNote · V2EX
- n=97
Gemma
▲ 2.6
67.8As usualexperience score
158 voices this week Confidence: High
- n=144
- Hacker News
- n=13
- Zhihu · RedNote · V2EX
- n=1
Gemma 4 31B is outperforming the family by 5.4 #1 in General text
MiniMax
–
63.7Better than usualexperience score
98 voices this week Confidence: High
- n=62
- Hacker News
- n=17
- Zhihu · RedNote · V2EX
- n=19
MiniMax H3 is outperforming the family by 8.1 #1 in Video generation
Qwen
▼ 0.4
61.3Worse than usualexperience score
1,232 voices this week Confidence: High
- n=950
- Hacker News
- n=142
- Zhihu · RedNote · V2EX
- n=140
Qwen3.6 27B is outperforming the family by 11.0 #1 in Local deploy
Kimi
▼ 1.2
59.5As usualexperience score
284 voices this week Confidence: High
- n=172
- Hacker News
- n=20
- Zhihu · RedNote · V2EX
- n=92
DeepSeek
▼ 0.4
49.1Dumber than usualexperience score
1,947 voices this week Confidence: High
- n=611
- Hacker News
- n=100
- Zhihu · RedNote · V2EX
- n=1,236
DeepSeek V4 Flash is outperforming the family by 23.6
Mistral
▲ 0.9
41.1Worse than usualexperience score
35 voices this week Confidence: High
- n=24
- Hacker News
- n=11
- Zhihu · RedNote · V2EX
- n=0
Best by task right now