GLM-5.2 emerges as the strongest all-round open-weight model of the cycle
As of early August, GLM-5.2 is rated the strongest all-round open-weight LLM — MIT-licensed, million-token context — while Kimi K2.7 Code leads open coding agents, Gemma 4 12B serves as a practical laptop model, and Nemotron 3 Super suits teams wanting open training resources. The open field has a clear all-rounder and clear specialists.
An all-round leader emerging is a sign of maturity. Early open-weight competition was a scramble of single benchmarks; a model recognised as the best general-purpose open option means the field now has a default foundation, the way the closed labs have a flagship. GLM-5.2 under MIT with long context is that default — capable and freely deployable.
The specialists around it matter as much as the leader. Kimi K2.7 Code for agentic coding, Gemma 4 12B for laptop-scale local use, Nemotron 3 Super for teams that want open training data — the open ecosystem now offers a fit-for-purpose model per workload, which is what a mature market looks like from the buyer's side.
The competitive pressure on closed labs is the through-line. When the best open all-rounder is MIT-licensed and the specialists cover coding, local, and training-transparent use cases, the closed premium has to justify itself workload by workload. The open field is no longer a fallback; it is a full toolbox, and the licences make it deployable.
Techsy — Best open-source LLMs: July 2026 leaderboard (updated) → · Thunder Compute — Best open source LLMs (August 2026) →