Granite 4.2 8B vs GLM-4.5-Air
Granite 4.2 8B scores higher on the intelligence index, 12 against 11. The cheapest machine here that runs GLM-4.5-Air is the Framework Desktop, 128GB, at $3,449. Granite 4.2 8B runs on the Mac mini M6, 24GB at $1,099, $2,350 less. On the Framework Desktop, 128GB, the cheapest machine here that runs both, Granite 4.2 8B is about 1.8× quicker: 18 tok/s against 10, both estimated from memory bandwidth. At 500k tokens a day the Framework Desktop, 128GB pays for itself in 139 years running GLM-4.5-Air, against 432 years running Granite 4.2 8B.
| Granite 4.2 8B | GLM-4.5-Air | |
|---|---|---|
| Intelligence index | 12 | 11 |
| Class | Below every hosted tier | Below every hosted tier |
| Weights | 5.3 GB | 73 GB |
| Needs at 32k | 11 GB | 79 GB |
| Quantisation | Q4_K_M | Q4_K_M |
| Parameters | 8B | 110.5B (12B active) |
| Max context | 128k | 128k |
| API price per 1M | $0.06 in / $0.15 out | $0.13 in / $0.85 out |
| Licence | Apache 2.0 | MIT |
| Machines here that run it | 33 of 37 | 12 of 37 |
| Cheapest machine that runs it | Mac mini M6, 24GB $1,099 | Strix Halo Framework Desktop, 128GB $3,449 |
Run Granite 4.2 8B on the Mac mini M6, 24GB · or GLM-4.5-Air on the Framework Desktop, 128GB
Jump to: Side by side on the Framework Desktop, 128GB · How much use it takes to pay for the machine · Machines that run one and not the other · The assumptions behind both columns
Side by side on the Framework Desktop, 128GB
The table above gives each model the cheapest machine that runs it, and those are two different machines, so nothing in it is a like-for-like race. The Strix Halo Framework Desktop, 128GB is the cheapest machine here that runs both, so this is the pair doing the same work on the same hardware.
| Granite 4.2 8B | GLM-4.5-Air | |
|---|---|---|
| Speed at 32k | 18 tok/s estimated | 10 tok/s estimated |
| Pay-back on this machine | Pays back in 432 years | Pays back in 139 years |
| API cost per month | $1.00 | $2.66 |
Run Granite 4.2 8B on the Framework Desktop, 128GB · or GLM-4.5-Air
How much use it takes to pay for the machine
Everything above is at 500k tokens a day. Pay-back moves with how much you actually run, so here are both models at the five levels of use the calculator names, on the Framework Desktop, 128GB. GLM-4.5-Air pays for it sooner at every level of use, so which of them to run does not turn on how hard you work it.
| A day's use | Granite 4.2 8B | GLM-4.5-Air |
|---|---|---|
| 50ka few chats a day | 4,321 years | 1,392 years |
| 200klight assistant use | 1,080 years | 348 years |
| 1Ma moderate coding-assistant day | 216 years | 70 years |
| 4Mheavy coding with an agent | 54 years | 17 years |
| 20Magents running most of the day | 11 years | 5.0 yearsits ceiling |
The Framework Desktop, 128GB generates at most 13.8M tokens a day on GLM-4.5-Air, so that column's figure at 20M tokens a day is for the most it can do, not for the whole of what was asked.
Run Granite 4.2 8B at 20M tokens a day · or GLM-4.5-Air
Machines that run one and not the other
Granite 4.2 8B needs 11 GB of memory at 32k of context and GLM-4.5-Air needs 79 GB. That puts Granite 4.2 8B on 21 of the 37 machines priced here that GLM-4.5-Air does not, starting at $1,099.
| Machine | Price | Memory | Speed on Granite 4.2 8B | Pay-back |
|---|---|---|---|---|
| Mac mini M6, 24GB | $1,099 | 24 GB | 12 tok/s estimated | Pays back in 122 years |
| Strix Halo Framework Desktop, 32GB | $1,269 | 32 GB | 18 tok/s estimated | Pays back in 159 years |
| Mac mini M6, 32GB | $1,299 | 32 GB | 12 tok/s estimated | Pays back in 144 years |
| AMD Radeon AI PRO R9700, 32GB | $1,299card only | 32 GB | 52 tok/s estimated | Pays back in 146 years |
| Mac mini M5 Pro, 24GB | $1,699 | 24 GB | 21 tok/s estimated | Pays back in 201 years |
| Strix Halo Corsair AI Workstation 300, 64GB | $1,700 | 64 GB | 18 tok/s estimated | Pays back in 213 years |
15 more, on the Granite 4.2 8B page.
Granite 4.2 8B is also head to head with Ling 3.0 tiny above it on the leaderboard. One more model needs much the same memory: Qwen3 8B. GLM-4.5-Air is also head to head with Mistral Small 4 (119B-2603) below it on the leaderboard. Two more models need much the same memory: Qwen3.5 122B-A10B and Devstral 2 123B.
The assumptions behind both columns
Both columns use the same usage: 500k tokens a day at 15:1 input to output, 32k of context, $0.17 per kWh, and today's API prices held flat. Speeds marked estimated are worked out from memory bandwidth rather than measured, and pay-back scales with them. Where nobody rents an open model by the token, its API prices are the nearest hosted model's, named beside them. Machines are the 37 here with a published price that are still sold. Change any of it in the calculator.
More head to head: every machine that runs Granite 4.2 8B · every machine that runs GLM-4.5-Air · every other match-up · both against the frontier · the quickest pay-back at each level of use