GPT

52.1 As usual experience score ▼ 0.8 Confidence: High

1,890 opinions tracked this week · 7-day rolling window

30-day experience score

Version breakdown

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

GPT-5.6 SolView lifecycle →	47.2−4.9	n=130	Confidence: High
GPT-5.4 Image 2	23.7−28.4	n=49	Confidence: High
GPT-5.6 LunaView lifecycle →	68.9+16.8	n=19	Confidence: Medium
GPT-5.5	64.5+12.4	n=7	Confidence: Low
GPT-5.6 Sol Pro	65.3+13.2	n=5	Confidence: Low

1,687 generic family mentions are included in the family score, not any version score.

By experience dimension

General text	53.9	n=759
Coding	59.0	n=391
Reasoning	60.2	n=277
Image / vision	37.4	n=108
Video generation	not enough data	n=1
Roleplay / creative	61.7	n=85
Safety & refusals	15.8	n=74
Speed & latency	40.3	n=211
Local deploy	not enough data	n=1

Top positive drivers

problem solved General text ×131
strong reasoning Reasoning ×120
problem solved Coding ×59
competitor flight General text ×59
speed quality Speed & latency ×54

Top negative drivers

performance issue Speed & latency ×78
safety overreach Safety & refusals ×45
hallucination Reasoning ×40
creativity down General text ×40
performance issue General text ×40

What users are saying

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

🔴 “GPT ok but requires significant prompt tuning.”
via Hacker News · General text · 2026-08-26

🟢 “Claude's default writing style tends to be more verbose than GPT's. GPT models are generally more concise when coding, while Claude often explains more.”
via Hacker News · Coding · 2026-08-26

🟢 “In April 2025, I used chatgpt again and then Gemini and was absolutely floored by the progress.”
via Reddit · Reasoning · 2026-08-26

🔴 “Claude and ChatGPT do this all the time: I tell them to change X in a planning document or email draft, and instead of just changing X they also frequently add the edit instruction to "change X" into the document itself.”
via Hacker News · Reasoning · 2026-08-26