The best local LLM for an iMac M6 (32 GB) is Qwen 3.6-35B-A3B at 59 tok/s (community run), the top-ranked model that reaches 15 tok/s here. Its 32 GB fits 63 of the 130 models in the LLMCheck index, up to 39.8B-class. The most capable, Qwen3.8-27B, runs at 10 tok/s (community run): slow for chat.
Considering this Mac?
iMac M6 on Amazon → · not sure? the Mac Advisor picks a config →
Ranked by LLMCheck suitability (capability balanced against speed on the M6); the top pick is the best-ranked model that reaches 15 tok/s here, has a standard Mac runtime and fits the GPU memory limit (raised if need be), and slower models are tagged slow. Click a model for its full figures and setup. Every speed carries its provenance mark: most are index estimates from the memory-bandwidth model, a filled square is a published figure and a hollow circle a community run — submit a real run →
| # | Model | Size | License | Speed | 64K context | Capability |
|---|---|---|---|---|---|---|
| 1 | Qwen 3.6-35B-A3B | 35B MoE | Apache 2.0 | community59 tok/s | agent-ready | 38/50 |
| 2 | Qwen3.8-27B slow | 27.8B | Apache 2.0 | community10 tok/s | agent-ready | 46/50 |
| 3 | Nanbeige4.3-33B slow | 33.1B | Apache 2.0 | estimated7 tok/s | 46/50 | |
| 4 | Qwen 3.6-27B slow | 27B | Apache 2.0 | estimated9 tok/s | agent-ready | 44/50 |
| 5 | KAT-Coder-V2.5 | 35B MoE | Apache 2.0 | estimated47 tok/s | agent-ready | 39/50 |
| 6 | LFM2.5-24B-A2B | 24B MoE | LFM Open | estimated79 tok/s | agent-ready | 35/50 |
| 7 | Recursive Flint | 39.8B MoE | Apache 2.0 | estimated61 tok/s | agent-ready | 37/50 |
| 8 | Muse Glimmer 30B slow | 30B | Apache 2.0 | community9 tok/s | agent-ready | 42/50 |
| 9 | MiniCPM-V 4.7 35B-A3B | 35.2B MoE | Apache 2.0 | estimated54 tok/s | agent-ready | 35/50 |
| 10 | Nemotron 3.5 Lightning | 30B MoE | OpenMDW | community45 tok/s | agent-ready | 36/50 |
| 11 | Gemma 4 31B slow | 31B | Apache 2.0 | community8 tok/s | 40/50 | |
| 12 | Raon-1 mini preview | 28.8B MoE | Apache 2.0 | estimated58 tok/s | agent-ready | 34/50 |
Showing the top 12 of 63 models that fit in 32 GB. See the full leaderboard or all benchmarks.
Context: “fits” means the weights at a short prompt. Coding agents (Claude Code, OpenCode, Codex CLI) hold 64K tokens or more, and every token adds KV cache: agent-ready marks a model whose weights plus a 64K-token 16-bit KV cache still fit in the 28 GB this Mac’s GPU can use with the limit raised (51 of the 58 models here with a KV figure). That is an estimate from each model’s config.json, not a measurement; a q8 KV cache halves the cache part, and — means the index has no KV figure for the model yet. The GPU memory limit below applies on top.
GPU memory limit: macOS lets the GPU use about 25 GB of this Mac’s 32 GB by default (macOS 26 and later, the only macOS it runs). All 63 models that fit stay within it at a short prompt; a long context adds its KV cache on top. sudo sysctl iogpu.wired_limit_mb=28672 raises it to 28 GB, leaving 4 GB for macOS, until the Mac restarts (sudo sysctl iogpu.wired_limit_mb=0 restores the default). How the limit is read →
The fastest way to get started is Ollama. Install it, then pull the model by its tag (checked on ollama.com, 2026-11-14):
ollama run qwen3.6:35b-a3b · model card →
Prefer a GUI? LM Studio gives you a one-click download and chat window. For step-by-step help see our Ollama install guide, or open the Qwen 3.6-35B-A3B on M6 page for its figures and setup.
Twice as fast: the cheapest Mac in the index that runs Qwen 3.6-35B-A3B at 2× or more of this Mac’s community59 tok/s is the Mac Studio M5 Max (48 GB) at community130 tok/s, $3,099 new (Apple list price, checked 2026-11-14): build-to-order from Apple (Amazon does not stock this configuration).
For less: the cheapest Mac that still runs Qwen 3.6-35B-A3B at 15+ tok/s is the MacBook Air M2 (24 GB) at estimated~43 tok/s, about $850 refurbished or used: search Amazon Renewed listings →. New: the Mac mini M6 (24 GB) at community59 tok/s, $1,099 new (Apple list price, checked 2026-11-14): Mac mini M6 (24GB) on Amazon →.
Speeds are each Mac’s own figure from the index, with its provenance mark. Prices come from data/macs.json: Apple’s list price for a new Mac (dated), an approximate street price for a used or refurbished one. Amazon stock checked 2026-11-18. Amazon links are affiliate links: as an Amazon Associate, LLMCheck earns from qualifying purchases.
The iMac M6 (32 GB) runs 63 of the models in the LLMCheck index, led by Qwen 3.6-35B-A3B. It is sold new by Apple; Amazon stock of the 32 GB configuration is not verified, so check the listing’s memory:
As an Amazon Associate, LLMCheck earns from qualifying purchases. Affiliate links cost you nothing extra and never influence our rankings.
LLMCheck. (2026). Best Local LLMs for the iMac M6 (32 GB). llmcheck.net. https://llmcheck.net/best-llm/imac-m6-32gb/ (updated ). CC BY 4.0.
Qwen 3.6-35B-A3B (35B MoE, Apache 2.0) is the best all-round pick at 59 tok/s (community run) on the iMac M6. If you want maximum speed, Ling-3.0-tiny hits ~132 tok/s (estimated); Qwen3.8-27B is more capable but runs at only 10 tok/s (community run) here.
About 63 of the 130 models in the LLMCheck index fit in 32 GB of unified memory, from compact models up to Recursive Flint (39.8B MoE).
No. A dense 70B such as Llama 3.3 70B needs about 44 GB at 4-bit; with 32 GB, stick to models like Qwen 3.6-35B-A3B. For a 70B, look at a 64 GB+ Mac.
32 GB is enough for dense models up to about 33B at 4-bit (the largest in the index that fits is Nanbeige4.3-33B), and for mixture-of-experts models such as KAT-Coder-V2.5 (~47 tok/s (estimated)). As a rule of thumb a 4-bit model needs about 0.57 GB per billion parameters plus ~2 GB, and macOS needs its own share of memory too. Because Apple Silicon uses unified memory, the same pool serves as system RAM and GPU memory.