Hands-on: GPT-6 Sol and Luna feel like improved 5.6, the real upgrade is pricing
Miyamoto_-_Musashi · reddit · 2026-09-23
The author argues people are overcomplicating GPT-6's Sol and Luna. After real work use and benchmark checks, he finds they feel mostly like improved versions of 5.6, with the biggest upgrade being the price. He notes Luna already followed this pattern before: small improvements plus better pricing, without feeling like a whole new model.
More from Models
- antirez backs low-bit DeepSeek 4.1 Flash over Qwen 27B dense on Apple M5 Macs — antirez · 2026-11-12
- Gemini re-renders images as text when vision is removed, dropping only 6% WSR — zacharynado · 2026-11-12
- Bare-harness coding benchmark: Fable 5.2 best overall, Gemini 4 Carbon best value — Semantic_meaning · 2026-11-12
- Google's Physical-Engineering's Last Exam: all 14 frontier models score below 50% — dl_weekly · 2026-11-12
- Reddit's 2026 LLM hall of fame: Opus 5.5 still unbeaten after 9 rival releases — G0dZylla · 2026-11-12
- Commit leak reveals unreleased Gemini 4 variants: Argon, Hydrogen and Helix (unconfirmed) — BlackHC · 2026-11-11