Qwen3 30B-A3B Instruct 2507 vs Laguna XS 2.1
Qwen3 30B-A3B Instruct 2507 scores 10 on the intelligence index, and Laguna XS 2.1 has not been placed on it. Both take the same machine to start: the cheapest here that runs either is the Framework Desktop, 32GB, at $1,269. On it, Laguna XS 2.1 is about 1.6× quicker: 44 tok/s against 27, both estimated from memory bandwidth. At 500k tokens a day the Framework Desktop, 32GB pays for itself in 78 years running Qwen3 30B-A3B Instruct 2507, against 127 years running Laguna XS 2.1.
| Qwen3 30B-A3B Instruct 2507 | Laguna XS 2.1 | |
|---|---|---|
| Intelligence index | 10 | not placed |
| Class | Below every hosted tier | not yet placed |
| Weights | 19 GB | 20 GB |
| Needs at 32k | 22 GB | 22 GB |
| Quantisation | Q4_K_M | Q4_K_M |
| Parameters | 30.5B (3.3B active) | 33.4B (3B active) |
| Max context | 256k | 256k |
| API price per 1M | $0.09 in / $0.3 out | $0.06 in / $0.12 out |
| Licence | Apache 2.0 | OpenMDW 1.1 |
| Machines here that run it | 28 of 37 | 28 of 37 |
| Cheapest machine that runs it | Strix Halo Framework Desktop, 32GB $1,269 | Strix Halo Framework Desktop, 32GB $1,269 |
Run Qwen3 30B-A3B Instruct 2507 on the Framework Desktop, 32GB · or Laguna XS 2.1
Jump to: Side by side on the Framework Desktop, 32GB · How much use it takes to pay for the machine · Memory is not what separates them · The assumptions behind both columns
Side by side on the Framework Desktop, 32GB
The cheapest machine that runs either model is the same one, so this is the pair doing the same work on the same hardware: Strix Halo Framework Desktop, 32GB, at $1,269.
| Qwen3 30B-A3B Instruct 2507 | Laguna XS 2.1 | |
|---|---|---|
| Speed at 32k | 27 tok/s estimated | 44 tok/s estimated |
| Pay-back on this machine | Pays back in 78 years | Pays back in 127 years |
| API cost per month | $1.57 | $0.97 |
Run Qwen3 30B-A3B Instruct 2507 on the Framework Desktop, 32GB · or Laguna XS 2.1
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, 32GB. Qwen3 30B-A3B Instruct 2507 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 | Qwen3 30B-A3B Instruct 2507 | Laguna XS 2.1 |
|---|---|---|
| 50ka few chats a day | 785 years | 1,269 years |
| 200klight assistant use | 196 years | 317 years |
| 1Ma moderate coding-assistant day | 39 years | 63 years |
| 4Mheavy coding with an agent | 9.8 years | 16 years |
| 20Magents running most of the day | 24 months | 3.2 years |
Run Qwen3 30B-A3B Instruct 2507 at 20M tokens a day · or Laguna XS 2.1
Memory is not what separates them
Qwen3 30B-A3B Instruct 2507 needs 22 GB of memory at 32k of context and Laguna XS 2.1 needs 22 GB. Every machine priced here that runs one runs the other, at 32k of context. So the choice between them is what each is good at, how fast it runs and what the same work costs on an API, not what you have to buy to hold it.
Qwen3 30B-A3B Instruct 2507 is also head to head with Mistral Small 4 (119B-2603) above it on the leaderboard and Qwen3-Coder 30B-A3B, the current Qwen nearest it in size. Laguna XS 2.1 is also head to head with Qwen3-Coder 30B-A3B and Hy-Memory 30B-A3B, which need much the same memory.
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 Qwen3 30B-A3B Instruct 2507 · every machine that runs Laguna XS 2.1 · every other match-up · both against the frontier · the quickest pay-back at each level of use