Ling

51.9 Dumber than usual experience score ▲ 1.3 Confidence: Medium

26 opinions tracked this week · 7-day rolling window

30-day experience score

405060708007-2908-0508-1308-2008-27

Version breakdown

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

Ling-3.0-flash 43.2−8.8 n=10 Confidence: Medium

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

By experience dimension

General text 47.4 n=5
Coding not enough data n=3
Reasoning not enough data n=3
Image / vision not enough data n=0
Video generation not enough data n=0
Roleplay / creative not enough data n=1
Safety & refusals not enough data n=0
Speed & latency 60.0 n=14
Local deploy not enough data n=0

Top positive drivers

  • speed quality Speed & latency ×7
  • strong reasoning Reasoning ×1
  • surprise impressive Reasoning ×1
  • aux model context compression General text ×1
  • release rollout General text ×1

Top negative drivers

  • speed quality Speed & latency ×2
  • performance issue Coding ×1
  • tool failure Coding ×1
  • hallucination Reasoning ×1
  • emotional companionship Roleplay / creative ×1

What users are saying

Selected from the latest 7 days, prioritizing recency, evidence quality, confidence, and engagement.

⚪ “Ling Q4 ran around 35TPS without any optimization on my halo strix, so I expect similar”

via Reddit · Speed & latency · 2026-08-25

🟢 “I'm playing with Ling 3.0 Tiny and is really enjoyable”

via Reddit · Reasoning · 2026-08-24

🟢 “Trying Ling 3.0 Tiny rn. Looks like a little beast, flying at 60 Tok/sec. Amazing! To me looks on par with Gemma 4 12b QAT but really faster.”

via Reddit · Speed & latency · 2026-08-24

⚪ “The model is not so good at coding too, comparing to qwen and gemma to be honest.”

via Reddit · Coding · 2026-08-24

🔴 “I tried Ling 3.0 Tiny at Q4_K_M. Thinking must be enabled or it will be very stupid. But enabling thinking makes it slower than e.g. gemma4-E4B or Qwen3.5-4B for simple tasks like translation or spam detection.”

via Reddit · Speed & latency · 2026-08-24

How's your AI experience today?

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