A static guide to local AI models that fit in a 128GB RAM budget. Built from the LocalClaw model database and ranked by hardware fit, use case, quality and speed.
With 128GB RAM, prioritize models with minimum RAM at or below 128GB and avoid filling memory completely. For most users, start with Xing4.0-29B-A4B, then test a faster smaller model if latency matters.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Official China Telecom XingChen-AGI Apache 2.0 MoE release with 29B total parameters, 4B active parameters, 256K native context and an official IQ4_NL GGUF path for llama.cpp-class local inference on 24GB GPU workstations.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Accio Lab Apache 2.0 co-work model post-trained from Qwen3.6-35B-A3B for long-horizon tools, files, code and business workflows. Official GGUF Q4_K_M is 19.7GiB with llama.cpp validation evidence.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Official Nex-AGI Apache 2.0 multimodal agent model for computer use, web browsing, coding and tool calling. Community Q4_K_M GGUF is about 21.3GB with documented llama.cpp text and vision smoke tests.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Official Qwen sparse multimodal MoE preview with 125B model parameters plus 51B n-gram embeddings, about 6B active parameters, Qwen Community 1.0 licensing, 262K native context and local Q4_K_M GGUF paths for llama.cpp, Ollama and LM Studio-class runtimes.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Official MIT 35B MoE reasoning model from Ornith AI with about 3B active parameters, 262K context, strong agentic-coding positioning and official Q4_K_M GGUF plus MLX/Ollama/llama.cpp local paths for larger workstations.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Meta Superintelligence Lab local agent model with text+image input, 131K context, Apache 2.0 weights and official GGUF/ExecuTorch artifacts. The K-Quant 17GB build targets 24GB machines; 32GB is safer for vision and long-context sessions.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Official Qwen dense 27B vision-language release with Apache 2.0 weights, 262K native context, thinking controls and strong agentic coding benchmarks. Practical local path through Unsloth and LM Studio-compatible GGUF artifacts.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: IBM Granite 4.2 30B brings the permissive Apache 2.0 Granite stack to workstation-class local reasoning, RAG, coding and tool-use workflows with GGUF and MLX community artifacts.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: IBM Granite 4.2 8B instruct model with Apache 2.0 weights, 128K context, thinking-mode chat template, tool calling and practical GGUF plus MLX paths for everyday local machines.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Large MoE model with only 10B active params. 60% cheaper to run than Qwen3-Max. 256K context. Top-tier reasoning, coding and multilingual. Hybrid think/non-think. Apache 2.0.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Qwen 3 dense 32B open-weight model with hybrid thinking and non-thinking modes, strong reasoning and coding support, and a practical Q4_K_M GGUF path for 32GB-class local machines.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Official MIT reasoning model from Ornith AI with a 262K native context window, tool-calling focus, and official Q4_K_M GGUF, MLX and Ollama/llama.cpp paths for local coding-agent experiments on 16GB+ machines.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: NVIDIA OpenMDW-1.1 hybrid Mamba/MoE/attention model for local agentic inference. The official GGUF path from ggml-org includes a 18.9GB Q4_0 build plus Ollama, llama.cpp and LM Studio recipes, with local contexts scaling from 4K to 256K+ depending on VRAM.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: InclusionAI's MIT-licensed instruct MoE optimized for fast agent workloads. 104B total parameters, only 7.4B active, hybrid linear attention, 262K context and strong tool-use / multi-step execution with high token efficiency.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Qwen Team Apache 2.0 coding-agent MoE with 80B total parameters, 3B active parameters, 262K native context and official Q4_K_M GGUF files. Practical local use fits best on 64GB+ workstations with recent llama.cpp or LM Studio-compatible runtimes.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Alibaba's next-gen MoE with hybrid-gated DeltaNet attention. Only 3B active params , runs at dense 7B speed with 70B quality. 256K native context (extensible to 1M). Hybrid thinking mode. Apache 2.0.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: InternScience Apache 2.0 agentic VLM. 35B-A3B MoE, 262K context, strong long-horizon search/tool-use benchmarks and official Q4_K_M GGUF artifacts for local workstations.
Why it fits: Comfortable memory headroom. General match. Catalogue summary: Qwen Team open-weight MoE for agentic coding and multimodal work. 35B total / 3B active, 262K native context, Apache 2.0, and strong GGUF availability through Unsloth and LM Studio-compatible artifacts.
This contextual order uses LocalClaw catalogue quality, reasoning, coding and speed fields plus freshness and RAM fit. It is not the homepage LocalClaw score, not a standardized third-party benchmark and never includes community stars. Catalogue summaries may repeat upstream claims; verify them at the linked model repository.