Have Open Weights Surpassed Closed Source?
markjeffrey · x · 2026-07-19
He argues that the "frontier" of open-weight models has already outgrown the closed-source camp:
- Models like DeepSeek v4 Pro and Kimi K3 are released as open weights with massive parameter scales.
- In contrast, Claude Fable 5 and GPT-5.6 Sol remain closed-source, with the latter only available to government-approved clients.
- He concludes that the visible boundaries of open-weight models have already surpassed closed-source ones, making it worth watching if "closed" will gradually become synonymous with "lagging behind" in the future.
More from AGI Musings
- The Evolution of LLM Business Models: Selling Outcomes Over Tokens — yacineMTB · 2026-07-22
- Bindu Reddy says GPT-6 is coming soon, with Alibaba, DeepSeek and Kimi close behind — bindureddy · 2026-07-22
- Bindu Reddy says the industry still lacks a way to train 20T models and scale post-training RL — bindureddy · 2026-07-22
- Advanced AI Models Are Becoming Impossible to Plug and Play — emollick · 2026-07-22
- AI suggested a better composition, and that made one user uneasy — Sydde · 2026-07-22
- The Thimble and the Waterfall: AI's Data Bottleneck and Feedback Loops — dyamins · 2026-07-22