What people run, by hardware
Pick the kind of machine you have. You'll see the model sizes people report running on it, most reported first, with the speed they typically get. It tells you what works on hardware like yours. It doesn't tell you which model gives the best answers.
To check a specific model on your exact card, use the calculator. For coding, see the coding models by VRAM.
A Mac counts when its report names the product, or when only that product was sold with its chip and memory.
Apple Silicon 18-36 GB· MacBook Pro only
Mid-size Macs. Up to about 30B-class models at 4 bits. Memory speed 154 to 461 GB/s across the 11 devices with reports, so a typical figure here spans all of them. 17 single-device reports.
| # | Model | Reports | Gen t/s | PP t/s | With speculation |
|---|---|---|---|---|---|
| 1 | Qwen3.635B · 3B active Alibaba · mostly 4bit Gen 25 t/s (3 runs) · PP 39 t/s (1 run) · with speculation 116 t/s (2) | 5 | 25 3 runs, 3 devices | 39 1 run | 116 2 runs |
| 2 | Qwen3.827B Alibaba · mostly IQ4_XS Gen 8.6 t/s (1 run) · PP 22 t/s (1 run) · with speculation 55 t/s (3) | 5 | 8.6 1 run | 22 1 run | 55 3 runs |
| 3 | Ornith1.535B · 3B active mostly 4-bit Gen 108 t/s (1 run) | 1 | 108 1 run | no PP data | none |
| 4 | Qwen330B · 3B active Alibaba · mostly IQ3_XXS Gen 74 t/s (1 run) | 1 | 74 1 run | no PP data | none |
| 5 | LFM2.524B · 2.5B active Liquid AI · mostly Q4_K_M Gen 52 t/s (1 run) | 1 | 52 1 run | no PP data | none |
| 6 | Qwen38B Alibaba · mostly Q4_K_M Gen 25 t/s (1 run) | 1 | 25 1 run | no PP data | none |
| 7 | GLM-5.3 Flash320B · 18B active Zhipu AI · mostly FP8-E4M3 No generation data | 1 | no generation data | no PP data | none |
| 8 | Gemma 426B · 4B active Google DeepMind · mostly QAT No generation data | 1 | no generation data | no PP data | none |
| 9 | Qwen3-Next80B · 3B active Alibaba · mostly Q4_K_M No generation data | 1 | no generation data | no PP data | none |
Not in the table: 1 report doesn't say which model size ran. 1 ran on more than one device, which says nothing about what one of them does.
How this list works
Each row is one model size, because a 27B and a 397B from the same family are different choices. Rows are ordered by how many people report running that size on this hardware.
Generation speed is the median of runs on one device, serving one request, without speculative decoding. PP is the median prompt-processing speed from plain runs that reported it, so its run count can be smaller. Speculative decoding (MTP, a draft model, DFlash, n-gram lookup) can make the same model several times faster, so those runs get their own column. A report that doesn't make clear which kind of run it was counts in neither.
Hardware is grouped by how much it holds and how fast its memory is, since memory speed sets how fast a model writes. Click a model for every report behind its row.