Open Models Tack Toward the Frontier | Tomasz Tunguz
Open-weight models have repeatedly reached equivalency with closed frontier models, first with DeepSeek R1, then GLM-4.6, GLM-5.2, & Kimi K3. But open source has not materially led; closed models such as GPT-5.2-class systems & Opus continue to create step-change moments.
In short : Open-weight models have repeatedly reached equivalency with closed frontier models, first with DeepSeek R1, then GLM-4.6, GLM-5.2, & Kimi K3. But open source has not materially led; closed models such as GPT-5.2-class systems & Opus continue to create step-change moments. Like two sailboats in a marathon race, open & closed AI labs are tacking & jibing in San Francisco Bay. In 2023, closed source models led by an enormous margin on Chatbot Arena Elo1. Two years later, the DeepSeek R1 moment arrived, the open-source answer to the ChatGPT moment. The two boats raced side-by-side…
saved by
related reading
- The Substitution Wave in AItomtunguz.com
- Clouded Judgement 7.24.26 - The World Isn’t Zero Sumcloudedjudgement.substack.com
- Google "We Have No Moat, And Neither Does OpenAI"semianalysis.com
- Deepseek: The Quiet Giant Leading China’s AI Racechinatalk.media
- Open models in perpetual catch-up - by Nathan Lambertinterconnects.ai
- What comes next with open models - by Nathan Lambertinterconnects.ai
- Open models in perpetual catch-upinterconnects.ai
- The Geopolitics of Open Weightsmbi-deepdives.com
- Paper AI Tigersgleech.org
- Who’s Afraid of Chinese Models? – Stratechery by Ben Thompsonstratechery.com
- GLM-5.2 Is The New Best Open Modelthezvi.substack.com
- GLM-5.2 is the step change for open agentsinterconnects.ai