Open labs embrace continued post-training: GLM 5.3, Qwen3.8 27B show generational leaps without new base models
Daniel_H212 · reddit · 2026-08-14
A Reddit user observes that open labs are adopting continued post-training instead of releasing entirely new models each generation. Examples like GLM 5.3, Qwen3.8 27B, and Deepseek V4 Flash 0731 show generational performance gains from post-training existing models. This benefits the local community as inference engines can keep up. Closed labs like Anthropic and OpenAI have been doing this for a while.
More from Models
- 'Small-Claude-shame': telling Opus 4.7 it's OK to know who you may ease it — repligate · 2026-10-16
- Bridgewater fine-tunes Qwen3-235B with Tinker, beats GPT-5.5 at 1/14th the cost — simonguozirui · 2026-10-16
- ChatGPT voice selection breaks after GPT-Live-1.1 rollout on Android — shaunralston · 2026-10-16
- Musk teases two-way Grok x Optimus integration: Grok in the bot, bot in Grok — beffjezos · 2026-10-16
- vLLM officially supports Muse Spark Open: multimodal, 1M context, agentic — vllm_project · 2026-10-16
- Liquid AI ships LFM2.5-24B-A2B: 24B MoE with 2B active params for local agents — JosephJacks_ · 2026-10-16