The best local LLMs for the M4 Max 128GB

Qwen3.5 122B-A10B (122B (10B active)) fits comfortably in the M4 Max 128GB's memory at Q4_K_M, using about 77GB with 8K context and generating ~31 tokens/sec. It is the most capable model in our catalog that runs well on this card. It is also our top pick for coding, reasoning and working with images. It is also the largest model that runs fully on this GPU.

Qwen3.5 122B-A10B is the mid-size mixture-of-experts model from Qwen's February 2026 Qwen3.5 launch, with 10B active parameters, image input, a 262K context, and switchable thinking. Later Qwen releases refreshed the 27B and 35B-A3B sizes, so this remains Qwen's main model for 96GB to 128GB machines. It needs about 77GB at Q4_K_M with 8K context, which suits Macs with 96GB or more, Strix Halo 128GB systems, and DGX Spark; a 24GB GPU needs 96GB or more of system RAM to hold the offloaded experts. Its hybrid attention keeps the KV cache small, so long contexts are cheap.

Qwen3.6 35B-A3B (35B (3B active)) fits comfortably in the M4 Max 128GB's memory at Q8_0, using about 41GB with 8K context and generating ~51 tokens/sec. It is the next-best model that runs well here, and the one to try if the top pick doesn't suit you.

Qwen3.6 35B-A3B is the April 2026 refresh of Qwen's fast mixture-of-experts model, focused on agentic coding and on preserving its reasoning across multi-turn tool use. With only 3B active parameters it generates text very quickly, and its hybrid linear attention keeps the KV cache tiny: the full 262K context needs about 5GB. At Q4_K_M it needs about 24GB with 8K context, which is tight on a 24GB card and comfortable on 32GB GPUs and Macs with 36GB or more; 12GB to 16GB GPUs run it well with experts offloaded to system RAM. It replaces Qwen3 30B-A3B and Qwen3 Coder 30B-A3B for most local users.

gpt-oss-20b (20.9B (3.6B active)) fits comfortably in the M4 Max 128GB's memory at MXFP4, using about 15GB with 8K context and generating ~67 tokens/sec. Pick it when responsiveness matters more than squeezing out maximum quality.

gpt-oss-20b is OpenAI's small open-weight reasoning model, a 21B mixture-of-experts with 3.6B active parameters whose expert weights ship natively in 4-bit MXFP4. OpenAI has not released a successor, and it remains one of the best reasoning and tool-calling models that fits a 16GB GPU. The native weights take about 14GB, so 16GB cards and Macs with 24GB or more run it fully in memory, and 12GB cards run it with a little CPU offload. Reasoning effort is adjustable (low, medium, high) but cannot be turned off.