Best local LLM for RAG
Top 7 open-source picks for RAG and document retrieval, ranked by benchmark performance and real-world fit. Updated monthly.
Qwen 3 VL 30B-A3B
Qwen 3 VL's sweet spot: a 30B MoE with 3B active parameters and 256k context. Delivers most of the 235B's quality at a fraction of the hardware cost.
Nemotron Nano 3 30B-A3B
NVIDIA's Mamba-2 + Transformer hybrid MoE with 3B active out of 30B total parameters. A native 1M-token context with roughly 4× the throughput of Nemotron 2.
Qwen 3 30B-A3B
Alibaba's Qwen 3 MoE with 30B total and just 3B active parameters, supporting hybrid thinking mode. MMLU 81.4, AIME24 80.4, 100+ languages, Apache 2.0.
Trinity Mini 26B-A3B
Arcee AI's US-built MoE with 3B active parameters out of 26B total. Apache-licensed, fast in practice, and tuned for agent-style workloads.
Kanana 2 30B-A3B Thinking
Kakao's agentic 30B MoE (3B active) with native hybrid thinking and Korean-first training. Apache 2.0 with MLA attention and 131k context.
Qwen 3 Omni 30B-A3B
Alibaba's omni-modal 30B MoE (3B active) with streaming speech, 119-language ASR, and Apache 2.0 licensing. The most accessible truly omnimodal open model.
Granite 4.0 H-Small 32B-A9B
IBM's hybrid Mamba-2 + MoE model with 32B total and 9B active parameters, engineered to slash long-context memory use by roughly 70% versus comparable transformers under Apache 2.0.