Open-Source LLM(Open-Source LLM)

An open-source LLM is a large language model whose weights — and often fine-tuning recipes — are released under a permissive license (typically MIT or Apache 2.0). This permits commercial use, self-hosted inference, and custom fine-tuning.

As of 2026, the leading open-source models include DeepSeek V4, Gemma 4, Qwen 3.5, Llama 4, GLM-5, and Mistral Medium. The Meta-dominated landscape has fragmented into a multi-polar one: Chinese labs (DeepSeek, Moonshot AI, Zhipu AI, Alibaba), Google, and Mistral all hold meaningful share. The performance gap with closed-source frontier models has converged to within one point on benchmarks like SWE-Bench and HumanEval.