https://runlocal.sh/new Every model added to the runlocal catalog, with honest "will it run?" verdicts for real hardware. en Tue, 07 Jul 2026 00:00:00 GMT https://runlocal.sh/m/gemma4 https://runlocal.sh/m/gemma4 Tue, 07 Jul 2026 00:00:00 GMT Google's Gemma 3 successor: unified any-to-any with vision, 256K context, Apache-2.0. Sizes: 12B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/granite-3-3 https://runlocal.sh/m/granite-3-3 Sun, 28 Jun 2026 00:00:00 GMT Sizes: 8B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/qwen3-coder https://runlocal.sh/m/qwen3-coder Sun, 28 Jun 2026 00:00:00 GMT Sizes: 30B-A3B (MoE). Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/qwq https://runlocal.sh/m/qwq Sun, 28 Jun 2026 00:00:00 GMT Sizes: 32B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/smollm3 https://runlocal.sh/m/smollm3 Sun, 28 Jun 2026 00:00:00 GMT Sizes: 3B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/falcon3 https://runlocal.sh/m/falcon3 Fri, 26 Jun 2026 00:00:00 GMT Sizes: 7B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/granite-3-1 https://runlocal.sh/m/granite-3-1 Fri, 26 Jun 2026 00:00:00 GMT Sizes: 8B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/nemotron-mini https://runlocal.sh/m/nemotron-mini Fri, 26 Jun 2026 00:00:00 GMT Sizes: 4B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/olmo2 https://runlocal.sh/m/olmo2 Fri, 26 Jun 2026 00:00:00 GMT Sizes: 7B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/yi-coder https://runlocal.sh/m/yi-coder Fri, 26 Jun 2026 00:00:00 GMT Sizes: 9B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/codellama https://runlocal.sh/m/codellama Wed, 24 Jun 2026 00:00:00 GMT Meta's Llama-2-based code models. Still popular on Ollama for completion. Sizes: 7B, 13B, 34B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/codestral https://runlocal.sh/m/codestral Wed, 24 Jun 2026 00:00:00 GMT Mistral's 22B code model with strong fill-in-the-middle. 32K context. Sizes: 22B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/command-r https://runlocal.sh/m/command-r Wed, 24 Jun 2026 00:00:00 GMT Cohere's RAG/tool-use specialist. No GQA, so KV cache is heavy at long context. Sizes: 35B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/deepseek-r1 https://runlocal.sh/m/deepseek-r1 Wed, 24 Jun 2026 00:00:00 GMT Open reasoning model. Distilled into Qwen/Llama backbones you can actually run locally. Sizes: 1.5B, 7B, 8B (Llama), 14B, 32B, 70B (Llama). Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/gemma2 https://runlocal.sh/m/gemma2 Wed, 24 Jun 2026 00:00:00 GMT Google's efficient open models. The 9B is a standout for its size. 8K context. Sizes: 2B, 9B, 27B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/gemma3 https://runlocal.sh/m/gemma3 Wed, 24 Jun 2026 00:00:00 GMT Google's latest, with vision and 128K context. Strong quality per parameter. Sizes: 1B, 4B, 12B, 27B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/llama-3-1 https://runlocal.sh/m/llama-3-1 Wed, 24 Jun 2026 00:00:00 GMT Meta's workhorse general model. Reliable all-rounder with 128K context. Sizes: 8B, 70B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/llama-3-2 https://runlocal.sh/m/llama-3-2 Wed, 24 Jun 2026 00:00:00 GMT Tiny Llamas for edge + on-device. Fast, light, 128K context. Sizes: 1B, 3B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/llama-3-2-vision https://runlocal.sh/m/llama-3-2-vision Wed, 24 Jun 2026 00:00:00 GMT Multimodal Llama — image understanding plus chat. 11B fits a 16GB+ Mac. Sizes: 11B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/llama-3-3 https://runlocal.sh/m/llama-3-3 Wed, 24 Jun 2026 00:00:00 GMT 70B that rivals far larger models. The best open general model that fits a 64GB Mac. Sizes: 70B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/llava https://runlocal.sh/m/llava Wed, 24 Jun 2026 00:00:00 GMT Popular open vision-language model for image Q&A and captioning. Sizes: 7B, 13B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/mistral https://runlocal.sh/m/mistral Wed, 24 Jun 2026 00:00:00 GMT The classic fast 7B. Apache-licensed, runs anywhere. Sizes: 7B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/mistral-nemo https://runlocal.sh/m/mistral-nemo Wed, 24 Jun 2026 00:00:00 GMT 12B with 128K context, built with NVIDIA. Great mid-size all-rounder. Sizes: 12B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/mistral-small https://runlocal.sh/m/mistral-small Wed, 24 Jun 2026 00:00:00 GMT 24B that punches at 70B level with much faster generation. 32K context. Sizes: 24B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/mixtral https://runlocal.sh/m/mixtral Wed, 24 Jun 2026 00:00:00 GMT Sparse MoE: 47B total but only ~13B active per token, so it decodes fast — if it fits in memory. Sizes: 8x7B (MoE). Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/mxbai-embed-large https://runlocal.sh/m/mxbai-embed-large Wed, 24 Jun 2026 00:00:00 GMT Strong open embedding model for retrieval. Small and quick. Sizes: 335M. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/nomic-embed-text https://runlocal.sh/m/nomic-embed-text Wed, 24 Jun 2026 00:00:00 GMT Fast, high-quality text embeddings for RAG. Tiny — runs anywhere. Sizes: 137M. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/phi-3-5-mini https://runlocal.sh/m/phi-3-5-mini Wed, 24 Jun 2026 00:00:00 GMT 3.8B with 128K context. No GQA, so KV grows fast at long context. Sizes: 3.8B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/phi-4 https://runlocal.sh/m/phi-4 Wed, 24 Jun 2026 00:00:00 GMT Microsoft's 14B that reasons above its weight, especially at math. 16K context. Sizes: 14B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine? https://runlocal.sh/m/phi-4-mini https://runlocal.sh/m/phi-4-mini Wed, 24 Jun 2026 00:00:00 GMT 3.8B with 128K context — a capable little reasoner for light rigs. Sizes: 3.8B. Fit verdicts + tok/s for Apple Silicon, NVIDIA and AMD — will it run on your machine?