https://localclaw.io/new Recently released open-weight AI models verified for local use in the LocalClaw catalogue. en Tue, 08 Sep 2026 12:00:00 GMT 1440 https://localclaw.io/models/minicpm5-2b https://localclaw.io/models/minicpm5-2b Tue, 08 Sep 2026 12:00:00 GMT minicpm Official OpenBMB compact on-device LLM with Apache 2.0 licensing, 131K context, tool-calling and coding focus, plus official Q4_K_M GGUF, Ollama, llama.cpp, Docker and OpenClaw run paths. 2B · 4 GB minimum RAM · Q4_K_M · minicpm. https://localclaw.io/models/spark-x2-5-4b https://localclaw.io/models/spark-x2-5-4b Wed, 02 Sep 2026 12:00:00 GMT spark Spark-X2.5-4B is an Apache 2.0 compact general-purpose model from XHToken with a hybrid attention architecture, 1M-token native context, multilingual coverage and official GGUF artifacts for local llama.cpp, Ollama and LM Studio-compatible workflows. 4B · 16 GB minimum RAM · BF16 GGUF · spark. https://localclaw.io/models/spark-x2-5-1-7b https://localclaw.io/models/spark-x2-5-1-7b Wed, 02 Sep 2026 12:00:00 GMT spark Spark-X2.5-1.7B is the smaller Apache 2.0 Spark-X2.5 release, tuned for lightweight conversation, coding, reasoning and agentic workflows with a 1M-token native context claim and official GGUF local runtime artifacts. 1.7B · 8 GB minimum RAM · BF16 GGUF · spark. https://localclaw.io/models/ibnsina-1.5b https://localclaw.io/models/ibnsina-1.5b Tue, 01 Sep 2026 12:00:00 GMT ibnsina IbnSina-1.5B is a Persian-first 1.48B Llama-compatible language model trained from scratch on a Persian-heavy corpus, with Apache 2.0 weights and GGUF artifacts for laptop, phone, Ollama, LM Studio and llama.cpp use. 1.5B · 4 GB minimum RAM · Q4_K_M · ibnsina. https://localclaw.io/models/k2-horizon-0-9b https://localclaw.io/models/k2-horizon-0-9b Tue, 01 Sep 2026 12:00:00 GMT k2-horizon IFM Apache 2.0 compact dense K2 Horizon model with 128K context, multi-teacher distillation for math/code/STEM tasks, and an official BF16 GGUF path for llama.cpp-compatible local experiments. 0.9B · 8 GB minimum RAM · BF16 GGUF · k2-horizon. https://localclaw.io/models/k2-horizon-mova-36b-a4b https://localclaw.io/models/k2-horizon-mova-36b-a4b Tue, 01 Sep 2026 12:00:00 GMT k2-horizon IFM Apache 2.0 sparse K2 Horizon release with Mixture-of-Experts plus Mixture-of-Values attention, 36B total / 4B active parameters, 512K context and an official BF16 GGUF path for high-memory local workstations. 36B (4B active, MoE) · 96 GB minimum RAM · BF16 GGUF · k2-horizon. https://localclaw.io/models/deepseek-v4-flash-vision-exp https://localclaw.io/models/deepseek-v4-flash-vision-exp Mon, 31 Aug 2026 12:00:00 GMT deepseek-flash Official MIT DeepSeek V4 Flash multimodal experiment with image understanding, 1M context and Unsloth Dynamic GGUF artifacts. The lightest practical GGUF is roughly 82-97GB, while higher-quality Q4/Q8 builds are about 155-162GB, so this belongs on large-memory workstations. 284B (13B active, multimodal MoE) · 128 GB minimum RAM · UD-Q2_K_XL · deepseek-flash. https://localclaw.io/models/granite4.2-8b https://localclaw.io/models/granite4.2-8b Wed, 26 Aug 2026 12:00:00 GMT granite 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. 8.8B · 8 GB minimum RAM · Q4_K_M · granite. https://localclaw.io/models/granite4.2-30b https://localclaw.io/models/granite4.2-30b Wed, 26 Aug 2026 12:00:00 GMT granite 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. 29.3B · 32 GB minimum RAM · Q4_K_M · granite. https://localclaw.io/models/granite4.2-3b https://localclaw.io/models/granite4.2-3b Tue, 25 Aug 2026 12:00:00 GMT granite IBM Granite 4.2 3B is the compact Apache 2.0 Granite reasoning model with 128K native context, thinking-mode chat, tool calling and official GGUF artifacts for laptop-class local inference. 3B · 8 GB minimum RAM · Q4_K_M · granite. https://localclaw.io/models/glm-5.3-flash https://localclaw.io/models/glm-5.3-flash Tue, 25 Aug 2026 12:00:00 GMT glm Z.ai MIT-licensed GLM-5 refresh with 320B total / 18B active parameters, native multimodal support, hybrid sparse-linear attention and a 1M-token context. Unsloth Dynamic GGUF makes it technically local, but it remains workstation/server-class hardware. 320B (18B active, MoE) · 160 GB minimum RAM · UD-IQ2_XXS · glm. https://localclaw.io/models/ornith-1-5-9b https://localclaw.io/models/ornith-1-5-9b Tue, 18 Aug 2026 12:00:00 GMT ornith 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. 9B · 16 GB minimum RAM · Q4_K_M · ornith. https://localclaw.io/models/ornith-1-5-35b-a3b https://localclaw.io/models/ornith-1-5-35b-a3b Tue, 18 Aug 2026 12:00:00 GMT ornith 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. 35B (3B active, MoE) · 48 GB minimum RAM · Q4_K_M · ornith. https://localclaw.io/models/dfm-mimir https://localclaw.io/models/dfm-mimir Fri, 14 Aug 2026 12:00:00 GMT dfm DFM-Mimir is a Danish and English 1B-class HRM language model trained from scratch by Danish Foundation Models, with Apache 2.0 weights, permissible-data positioning and GGUF artifacts for lightweight local inference. 1.8B · 4 GB minimum RAM · Q8_0 · dfm. https://localclaw.io/models/llm-jp-4-33b-thinking https://localclaw.io/models/llm-jp-4-33b-thinking Fri, 14 Aug 2026 12:00:00 GMT llm-jp Official Apache 2.0 LLM-jp reasoning model with English/Japanese support, 65K GGUF context metadata and an official Q4_K_M GGUF path for local llama.cpp and LM Studio testing on 64GB+ workstations. 33B · 64 GB minimum RAM · Q4_K_M · llm-jp. https://localclaw.io/models/nemotron-3-5-lightning-30b-a3b https://localclaw.io/models/nemotron-3-5-lightning-30b-a3b Tue, 11 Aug 2026 12:00:00 GMT nemotron 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. 30B (3B active, MoE) · 48 GB minimum RAM · Q4_0 · nemotron. https://localclaw.io/models/lfm2-5-vl-3b https://localclaw.io/models/lfm2-5-vl-3b Tue, 11 Aug 2026 12:00:00 GMT lfm Liquid AI edge vision-language model with LFM2.5-2.6B backbone, SigLIP2 NaFlex vision encoder, 32K context, LFM 1.0 open weights and official GGUF plus llama.cpp and MLX runtime paths for local image chat and OCR. 3B multimodal · 8 GB minimum RAM · Q4_K_M + mmproj · lfm. https://localclaw.io/models/ling-3.0-tiny https://localclaw.io/models/ling-3.0-tiny Mon, 10 Aug 2026 12:00:00 GMT ling Official InclusionAI MIT hybrid-reasoning MoE with 131K context, 1.3B active parameters and a validated GGUF path through stock llama.cpp builds from the 2026-08-17 bailingmoe3 merge. 7.9B (1.3B active, MoE) · 8 GB minimum RAM · Q4_K_M · ling. https://localclaw.io/models/muse-glimmer-30b https://localclaw.io/models/muse-glimmer-30b Sun, 09 Aug 2026 12:00:00 GMT meta 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. 29.8B multimodal · 24 GB minimum RAM · K-Quant 17GB Q4_K_M · meta. https://localclaw.io/models/qwen3.8-27b https://localclaw.io/models/qwen3.8-27b Wed, 05 Aug 2026 12:00:00 GMT qwen 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. 27B · 32 GB minimum RAM · Q4_K_M · qwen.