https://localclaw.io/new Recently released open-weight AI models verified for local use in the LocalClaw catalogue. en Sun, 27 Sep 2026 12:00:00 GMT 1440 https://localclaw.io/models/xing4-0-29b-a4b https://localclaw.io/models/xing4-0-29b-a4b Thu, 17 Sep 2026 12:00:00 GMT xing 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. 29B (4B active, MoE) · 32 GB minimum RAM · IQ4_NL GGUF · xing. https://localclaw.io/models/bonsai-2-27b https://localclaw.io/models/bonsai-2-27b Thu, 17 Sep 2026 12:00:00 GMT bonsai PrismML Apache 2.0 ternary Qwen3.8-27B derivative with official GGUF and MLX paths. The PQ2_0 pack is 7.21GB, PTQ1_0 is 5.95GB, and custom llama.cpp/MLX kernels target laptop-class local inference. 27.36B (ternary) · 16 GB minimum RAM · PQ2_0 · bonsai. https://localclaw.io/models/needle-3 https://localclaw.io/models/needle-3 Wed, 16 Sep 2026 12:00:00 GMT needle Cactus Compute Apache 2.0 foundation model for tiny on-device tool calling, extraction and embeddings. Official `.cact` runtime artifacts run through the cactus-needle Python package, C API, browser/WASI and platform runners instead of stock GGUF or LM Studio. 121M laddered SAN · 1 GB minimum RAM · CQ2 .cact · needle. https://localclaw.io/models/occamy-1-0 https://localclaw.io/models/occamy-1-0 Tue, 15 Sep 2026 12:00:00 GMT occamy 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. 35B (3B active, MoE) · 32 GB minimum RAM · Q4_K_M · occamy. https://localclaw.io/models/nex-n2-5-mini https://localclaw.io/models/nex-n2-5-mini Thu, 10 Sep 2026 12:00:00 GMT nex 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. 35B MoE · 32 GB minimum RAM · Q4_K_M · nex. https://localclaw.io/models/deepseek-v4-1-flash https://localclaw.io/models/deepseek-v4-1-flash Thu, 10 Sep 2026 12:00:00 GMT deepseek-flash Official MIT DeepSeek V4.1 Flash release with CED architecture, CSA2 attention, multimodal input and 1M-token context. Community GGUF artifacts include split Q2_K/Q3/Q4 builds plus a llama.cpp patch path; use only on 256GB+ workstations, with 384GB+ safer. 552B MoE (8B/16B active) · 256 GB minimum RAM · Q2_K · deepseek-flash. 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-3-7b https://localclaw.io/models/k2-horizon-3-7b Tue, 01 Sep 2026 12:00:00 GMT k2-horizon IFM Apache 2.0 dense K2 Horizon model with 3.7B parameters, 512K context, public training resources and an official BF16 GGUF artifact for K2 Horizon llama.cpp-compatible local experiments. 3.7B · 16 GB minimum RAM · BF16 GGUF · k2-horizon. https://localclaw.io/models/k2-horizon-7b https://localclaw.io/models/k2-horizon-7b Tue, 01 Sep 2026 12:00:00 GMT k2-horizon IFM Apache 2.0 dense K2 Horizon release with about 9B parameters, 128K context, open training-data references and an official BF16 GGUF artifact for K2 Horizon llama.cpp-compatible local experiments. 9B · 24 GB minimum RAM · BF16 GGUF · k2-horizon. https://localclaw.io/models/k2-horizon-32b https://localclaw.io/models/k2-horizon-32b Tue, 01 Sep 2026 12:00:00 GMT k2-horizon IFM Apache 2.0 dense K2 Horizon member with 32B parameters, 512K context, official Q4_K_M GGUF artifacts, and a documented local path through the K2 Horizon llama.cpp fork while upstream support matures. 32B dense · 32 GB minimum RAM · Q4_K_M · 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/qwen3.8-flash-next https://localclaw.io/models/qwen3.8-flash-next Wed, 26 Aug 2026 12:00:00 GMT qwen 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. 125B + 51B n-gram (6B active) · 96 GB minimum RAM · Q4_K_M · qwen. 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.