User reports 'thinking' mode no longer lengthens responses as before
Kittu_Mitthu · reddit · 2026-08-29
A Reddit user notes that back in July, enabling the thinking option made responses automatically longer without prompting. In a new chat today, the same option appears to have no effect, even after explicitly asking for longer paragraphs, leading them to suspect the behavior was changed.
More from Models
- Gemini 3.7 Flash Used 70% of Quota for a Single Coding Prompt — Mmd-NeWton · 2026-08-29
- antirez previews GLM 5.3 Flash update: vision, anti-refusal vectors, and ROCm support coming soon — antirez · 2026-08-29
- Users Complain Gemini Updates Make It Dumber: Biggest Downgrade Yet — guitarpluscoffee · 2026-08-29
- Leaked demos show DeepSeek Astra outputs at Max effort — Outside-Iron-8242 · 2026-08-29
- rednote explores open-weight multimodal model with 512K context for long-horizon agents — aftahi_ai · 2026-08-29
- GLM-5.3 outperforms Sol and Opus on new Terminal-bench-4.0 — nijfranck · 2026-08-29