Yes: Mellum2.1-12B-A2.5B-Thinking runs on 53 of the 54 Macs in the LLMCheck index. It needs about 10 GB at a 4K context, 11 GB at 64K. The smallest and cheapest current Mac that runs it at 15+ tok/s is the Mac mini M6 16 GB ($899; ~55 tok/s est.). Best measured figure: 112.7 tok/s on an M5 Pro (community).
One row per Mac configuration, grouped by product line. The build is the one each speed is for; the context column is the longest prompt that fits inside macOS’s default GPU memory limit with a 16-bit KV cache (“raised” where the limit has to be raised first; “full” when the whole window fits). Speeds are decode tok/s from the LLMCheck index, each with its provenance mark. For small mixture-of-experts models (40B total or less), the formula’s typical error is about 15% (the median gap to 88 measured plain-decode runs; it tends to undershoot (median −7%), and an estimate scaled from the model’s own measured run is usually closer; method).
| Mac | Verdict | Build | Speed | Max context | Model page |
|---|---|---|---|---|---|
| MacBook Air | |||||
| MacBook Air M5 16 GB | Runs | Q4_K_M | estimated55 tok/s | 128K (full) | — |
| MacBook Air M5 24 GB | Runs | Q4_K_M | estimated55 tok/s | 128K (full) | — |
| MacBook Air M5 32 GB | Runs | Q4_K_M | estimated55 tok/s | 128K (full) | — |
| MacBook Air M1 8 GB refurb | Doesn't fit | — | — | — | — |
| MacBook Air M4 16 GB used | Runs | Q4_K_M | estimated46 tok/s | 49K | M4 page |
| MacBook Air M2 16 GB refurb | Runs | Q4_K_M | estimated40 tok/s | 49K | — |
| MacBook Air M3 16 GB refurb | Runs | Q4_K_M | estimated40 tok/s | 49K | — |
| MacBook Air M1 16 GB refurb | Runs | Q4_K_M | estimated29 tok/s | 49K | — |
| MacBook Air M4 24 GB used | Runs | Q4_K_M | estimated46 tok/s | 128K (full) | M4 page |
| MacBook Air M2 24 GB refurb | Runs | Q4_K_M | estimated40 tok/s | 128K (full) | — |
| MacBook Air M3 24 GB refurb | Runs | Q4_K_M | estimated40 tok/s | 128K (full) | — |
| MacBook Pro | |||||
| MacBook Pro M6 16 GB | Runs | Q4_K_M | estimated55 tok/s | 128K (full) | — |
| MacBook Pro M5 Pro 24 GB | Runs | Q4_K_M | estimated128 tok/s | 128K (full) | M5 Pro page |
| MacBook Pro M6 24 GB | Runs | Q4_K_M | estimated60 tok/s | 128K (full) | — |
| MacBook Pro M6 32 GB | Runs | Q4_K_M | estimated60 tok/s | 128K (full) | — |
| MacBook Pro M5 Pro 48 GB | Runs | Q4_K_M | estimated128 tok/s | 128K (full) | M5 Pro page |
| MacBook Pro M5 Max 64 GB | Runs | Q4_K_M | estimated154 tok/s | 128K (full) | M5 Max page |
| MacBook Pro M5 Pro 64 GB | Runs | Q4_K_M | estimated128 tok/s | 128K (full) | M5 Pro page |
| MacBook Pro M5 Max 128 GB | Runs | Q4_K_M | estimated154 tok/s | 128K (full) | M5 Max page |
| MacBook Pro M3 Max 64 GB refurb | Runs | Q4_K_M | estimated140 tok/s | 128K (full) | — |
| MacBook Pro M3 Max 128 GB refurb | Runs | Q4_K_M | estimated140 tok/s | 128K (full) | — |
| Mac mini | |||||
| Mac mini M6 16 GB | Runs | Q4_K_M | estimated55 tok/s | 128K (full) | — |
| Mac mini M5 Pro 24 GB | Runs | Q4_K_M | estimated128 tok/s | 128K (full) | M5 Pro page |
| Mac mini M6 24 GB | Runs | Q4_K_M | estimated60 tok/s | 128K (full) | — |
| Mac mini M6 32 GB | Runs | Q4_K_M | estimated60 tok/s | 128K (full) | — |
| Mac mini M5 Pro 48 GB | Runs | Q4_K_M | estimated128 tok/s | 128K (full) | M5 Pro page |
| Mac mini M5 Pro 64 GB | Runs | Q4_K_M | estimated128 tok/s | 128K (full) | M5 Pro page |
| Mac mini M2 Pro 16 GB refurb | Runs | Q4_K_M | estimated67 tok/s | 49K | — |
| Mac mini M4 16 GB used | Runs | Q4_K_M | estimated46 tok/s | 49K | M4 page |
| Mac mini M4 Pro 24 GB used | Runs | Q4_K_M | estimated122 tok/s | 128K (full) | — |
| Mac mini M4 24 GB used | Runs | Q4_K_M | estimated46 tok/s | 128K (full) | M4 page |
| Mac mini M2 Pro 32 GB refurb | Runs | Q4_K_M | estimated67 tok/s | 128K (full) | — |
| Mac mini M4 Pro 48 GB used | Runs | Q4_K_M | estimated122 tok/s | 128K (full) | — |
| Mac mini M4 Pro 64 GB used | Runs | Q4_K_M | estimated122 tok/s | 128K (full) | — |
| iMac | |||||
| iMac M6 16 GB | Runs | Q4_K_M | estimated55 tok/s | 128K (full) | — |
| iMac M6 24 GB | Runs | Q4_K_M | estimated60 tok/s | 128K (full) | — |
| iMac M6 32 GB | Runs | Q4_K_M | estimated60 tok/s | 128K (full) | — |
| Mac Studio | |||||
| Mac Studio M5 Max 36 GB | Runs | Q4_K_M | estimated145 tok/s | 128K (full) | M5 Max page |
| Mac Studio M5 Max 48 GB | Runs | Q4_K_M | estimated154 tok/s | 128K (full) | M5 Max page |
| Mac Studio M5 Max 64 GB | Runs | Q4_K_M | estimated154 tok/s | 128K (full) | M5 Max page |
| Mac Studio M5 Ultra 96 GB | Runs | Q4_K_M | estimated174 tok/s | 128K (full) | — |
| Mac Studio M5 Max 128 GB | Runs | Q4_K_M | estimated154 tok/s | 128K (full) | M5 Max page |
| Mac Studio M5 Ultra 256 GB | Runs | Q4_K_M | estimated174 tok/s | 128K (full) | — |
| Mac Studio M5 Ultra 512 GB | Runs | Q4_K_M | estimated174 tok/s | 128K (full) | — |
| Mac Studio M2 Max 32 GB refurb | Runs | Q4_K_M | estimated140 tok/s | 128K (full) | — |
| Mac Studio M1 Ultra 64 GB refurb | Runs | Q4_K_M | estimated163 tok/s | 128K (full) | — |
| Mac Studio M4 Max 64 GB used | Runs | Q4_K_M | estimated150 tok/s | 128K (full) | — |
| Mac Studio M2 Max 64 GB refurb | Runs | Q4_K_M | estimated140 tok/s | 128K (full) | — |
| Mac Studio M1 Ultra 128 GB refurb | Runs | Q4_K_M | estimated163 tok/s | 128K (full) | — |
| Mac Studio M2 Ultra 128 GB refurb | Runs | Q4_K_M | estimated163 tok/s | 128K (full) | — |
| Mac Studio M4 Max 128 GB used | Runs | Q4_K_M | estimated150 tok/s | 128K (full) | — |
| Mac Studio M2 Ultra 192 GB refurb | Runs | Q4_K_M | estimated163 tok/s | 128K (full) | — |
| Mac Studio M3 Ultra 256 GB used | Runs | Q4_K_M | estimated164 tok/s | 128K (full) | — |
| Mac Studio M3 Ultra 512 GB refurb | Runs | Q4_K_M | estimated164 tok/s | 128K (full) | — |
Context memory is computed from its config.json (7 full layers + 21 sliding-window layers (window 1024)): about 1.0 GB of 16-bit KV cache at 64K tokens. A q8 cache (llama.cpp -ctk q8_0 -ctv q8_0, Ollama OLLAMA_KV_CACHE_TYPE=q8_0) halves that part. Runtimes also grow a compute buffer with context, so a fit within about 1 GB of the limit is optimistic.
macOS lets the GPU use only part of unified memory by default. Mellum2.1-12B-A2.5B-Thinking fits inside that default limit on every Mac it fits at all, so no Mac in the table needs it raised at a short context. Long contexts may still need it: the context column already uses each Mac’s default limit.
Weights are 12.15B parameters at each format’s effective size (the index’s GB-per-billion table), plus about 2 GB for the runtime and a short context; longer contexts add the KV cache computed from the model’s config.json at 16 bits. It is a mixture-of-experts model: all 12.15B parameters sit in memory although only 2.5B are read per token. The first row is the catalog’s own figure (about 10 GB at a 4K context), the one every verdict on this page uses. “Smallest Mac” is the smallest memory size among the Macs Apple sells new that holds it at a 4K context inside that Mac’s default GPU limit. Estimates, not measurements.
| Build | Weights | RAM at 4K | at 32K | at 64K | Smallest Mac |
|---|---|---|---|---|---|
| As catalogued | — | 10 GB | 10 GB | 11 GB | 16 GB |
| Q2_K | 4.1 GB | 6 GB | 7 GB | 7 GB | 16 GB |
| Q3_K_M | 6.0 GB | 8 GB | 8 GB | 9 GB | 16 GB |
| Q4_K_M | 6.9 GB | 9 GB | 9 GB | 10 GB | 16 GB |
| Q5_K_M | 8.6 GB | 11 GB | 11 GB | 12 GB | 16 GB |
| Q6_K | 10.0 GB | 12 GB | 12 GB | 13 GB | 24 GB |
| Q8_0 | 12.9 GB | 15 GB | 15 GB | 16 GB | 24 GB |
| F16 | 24.3 GB | 26 GB | 27 GB | 27 GB | 36 GB |
Every published (filled square) and community (hollow circle) run the index holds for Mellum2.1-12B-A2.5B-Thinking: 0 published, 4 community, each linked to its source. LLMCheck runs no tests of its own. One row is one measurement session; on each chip the table above shows the lower median of the eligible runs (plain decode at 5 bits or fewer), and a binned or “not an anchor” run is shown but never used as a reference.
| Chip | Runtime | Quant | Decode | Context | Date | Source |
|---|---|---|---|---|---|---|
| M5 Max | LM Studio | BF16 | community96.8 tok/s | short prompts, long generations | 2026-10-09 | llm-bench.io |
| M5 Pro | MLX | 6-bit | community112.7 tok/s | 1k | 2026-10-23 | omlx.ai |
| M4 | llama.cpp | Q4_K_M | community50 tok/s | short coding prompts | 2026-10-08 | note.com not an anchor |
| M4 | MLX | 6-bit | community35.9 tok/s | 4k | 2026-10-10 | omlx.ai |
The index has no verified Ollama tag for Mellum2.1-12B-A2.5B-Thinking. Start from its listing and copy the exact name of a build that fits: Hugging Face repo → · model card →
GGUF builds run in LM Studio or llama.cpp; MLX builds run in LM Studio or mlx-lm.
On the Mac mini M6 (16 GB), the smallest current Mac that runs it inside the default GPU limit, the index figure is ~55 tok/s (estimated) at Q4_K_M; keep the context at or under 128K there, or use a q8 KV cache.
Step-by-step: the Ollama install guide.
Mac mini M6 (16 GB) — smallest current Mac that runs it and cheapest at 15+ tok/s, $899 new (Apple list price, checked 2026-11-14), ~55 tok/s (estimated).
As an Amazon Associate, LLMCheck earns from qualifying purchases. Affiliate links cost you nothing extra and never influence the verdicts.
Yes. Mellum2.1-12B-A2.5B-Thinking needs about 10 GB of unified memory at a short (4K-token) context, so it runs on 53 of the 54 Mac configurations in the LLMCheck index. The smallest Mac Apple sells new that runs it is the Mac mini M6 (16 GB).
About 10 GB at a 4K context: the weights plus about 2 GB for the runtime. Context adds KV cache: about 11 GB in total at 32K and 11 GB at 64K with the default 16-bit cache (about 11 GB at 64K with a q8 cache). These are estimates from the model’s config.json, not measurements.
On an M5 Max, the chip the leaderboard quotes it on, the index figure is ~154 tok/s (estimated) at Q4_K_M. The best measured figure is 112.7 tok/s on an M5 Pro 64 GB (community run, 6-bit). Every Mac’s figure is in the table on this page with its provenance: published, community run or estimated.
Among Macs Apple sells new, the Mac mini M6 (16 GB), $899 new (Apple list price, checked 2026-11-14), is the cheapest that reaches 15 tok/s with Mellum2.1-12B-A2.5B-Thinking, at ~55 tok/s (estimated).
No. On all 53 Mac configurations where Mellum2.1-12B-A2.5B-Thinking fits, it fits inside that Mac’s default GPU memory limit at a short context; only a long context on the smallest of them can push it over.
Q4_K_M (about 7 GB of weights) is the common 4-bit GGUF choice. Q8_0 (about 13 GB) is closer to full quality and needs a 24 GB Mac at the default GPU limit. Q3_K_M and Q2_K shrink it to about 6 and 4 GB at a quality cost. The index’s reference figure (154 tok/s on an M5 Max) is for its Q4_K_M build.
Its own window is 128K tokens. At macOS’s default GPU limit, the Mac mini M6 (16 GB) holds about 128K (the full window).
LLMCheck. (2026). Can I run Mellum2.1-12B-A2.5B-Thinking on a Mac? llmcheck.net. https://llmcheck.net/can-i-run/mellum21-12b-a25b-thinking/ (updated ). CC BY 4.0.