Laguna M.1 — Hardware Requirements & GPU Compatibility
ChatFunctions
Laguna M.1 is a 225.8B-parameter open language model from poolside. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 136.07 GB of VRAM — see which GPUs and Macs can run it below.
Laguna M.1 runs on a 192 GB+ Mac or a datacenter GPU
Best local pick → Mac Pro M2 Ultra (192 GB) · Q4_K_M · 138.1 GB · ~26 tok/s
Macs from 96 GB (IQ3_XXS · ~33 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Laguna M.1
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
ollama run hf.co/sigargv/Laguna-M.1-GGUF:Q4_K_MQ4_K_M runs on one H200 NVL 141 GB · 16K context
Download: Q4_K_M/Laguna-M.1-Q4_K_M-imatrix-public-v1-00001-of-00010.gguf136.72 GB + 9 partssigargv
Other ways
each opens step-by-stepRun in cloud
Fits on B200 (192 GB) (53 GB headroom) · Q4_K_M
- Generation speed
- ~152 tok/s
- generation speed
- Cost per 1M output tokens
- $10.91
- per 1M output tokens
How Much VRAM Does Laguna M.1 Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | Runs on | |
|---|---|---|---|---|
| 68.3 GB | 72.7 GB | 2-bit · i-quant extra small — Importance-weighted 2-bit, extra small | Runs on RTX PRO 5000 Blackwell 72 GB | |
| 96.6 GB | 101 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Runs on Instinct MI250X 128 GB | |
| 99.4 GB | 103.8 GB | 3-bit small — 3-bit small quantization | Runs on Instinct MI250X 128 GB | |
| 110.7 GB | 115.1 GB | 3-bit medium — 3-bit medium quantization | Runs on Instinct MI250X 128 GB | |
| 136.1 GB | 140.5 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Runs on H200 NVL 141 GB | |
| 161.5 GB | 165.9 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Runs on B200 192 GB | |
| 186.9 GB | 191.3 GB | 6-bit, very good — 6-bit quantization, very good quality | Runs on B200 192 GB | |
| 226.4 GB | 230.8 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Runs on Instinct MI325X 256 GB |
Selected Q4_K_M
4-bit medium quantization138.1GBat 16K context
- Weights
- 135.5 GB
- KV cache
- 2.4 GB
- Overhead
- 0.3 GB
Server-class: ~138.1 GB at 16K context · one H200 NVL 141 GB
Which Hardware Can Run Laguna M.1?
Q4_K_M · 136.1 GBLaguna M.1 (Q4_K_M) requires 136.1 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 177+ GB is recommended. Using the full 262K context window can add up to 37.3 GB, bringing total usage to 173.4 GB. No consumer GPU has enough memory. 4 devices with unified memory can run Laguna M.1, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Pro M2 Ultra (192 GB).
128–144 GB
Q4_K_M · best fit IQ4_XS (124.0 GB)
4 GPUs · 2 Mac classes
GH200 Grace Hopper Superchip~128Simulate output on GH200 Grace Hopper SuperchipH200 NVL~126Simulate output on H200 NVLH200 SXM~126Simulate output on H200 SXMM2 Ultra*128 GB~27Simulate output on Mac Pro M2 Ultra (128 GB)M5 Max128 GB~23Simulate output on Mac Studio M5 Max (128 GB)192+ GB
fits · 53.9 GB free · up to 256K ctx
6 GPUs · 1 Mac class · 2 device families
B200~152Simulate output on B200B300~152Simulate output on B300Instinct MI350X~59Simulate output on Instinct MI350XM3 Ultra*256–512 GB~26Simulate output on Mac Studio (M3 Ultra, 256GB)8x H100 SXM5~215Simulate output on NVIDIA DGX H1008x A100 SXM4~193Simulate output on NVIDIA DGX A100 640GB
By VRAM tier · typical GPUs and Macs at Q4_K_M
Consumer GPUs (up to 32 GB) and Macs up to 64 GB: too small for this model (needs ≥ 70.4 GB)8 GB tier →12 GB tier →16 GB tier →24 GB tier →Macs →
How good is Laguna M.1?
llmrun Score blends public benchmark boards into one number; 100 = the average open model.
Not enough data yet for an llmrun Score.
A Score needs results on enough independent boards. How the Composite Score is built
Benchmarks
Here are all 3 boards that have results for this model, sorted by their weight in the llmrun Score; badges show who ran the test.
- LMArena WebDev
LMArena · Oct 2026Arena
Arena1349.7
#39/52 open
- Surface Evolver Bench
Epoch AI · Apr 2026Lab-reported
Lab-reported15.6%
#13/13 open
- ProofBench
Epoch AI · Apr 2026Lab-reported
Lab-reported0.0%
#24/26 open
Similar models
4| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| MiniMax M2.7MiniMax | 107.4 | 139.6 GB | —needs > 24 GB | same size |
| Command A Plus 05 2026 BF16Cohere | — | 132.1 GB | —needs > 24 GB | same size |
| Qwen3 VL 235B A22B InstructAlibaba | — | 143.3 GB | —needs > 24 GB | same size |
| DeepSeek V2.5 1210DeepSeek | — | 142.9 GB | —needs > 24 GB | same size |
Specifications
- Publisher
- poolside
- Parameters
- 225.8B
- Active parameters
- 23.5B per token
- Architecture
- LagunaForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 100,352
- Release Date
- 2026-06-15
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Laguna M.1 need?
Laguna M.1 requires 136.1 GB of VRAM at Q4_K_M, or 452.2 GB at BF16. Full 262K context adds up to 37.3 GB (173.4 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 225.8B × 4.8 bits ÷ 8 = 135.5 GB
KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 37.9 GB (at full 262K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Laguna M.1?
No — Laguna M.1 requires at least 68.3 GB at IQ2_XS, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
What's the best quantization for Laguna M.1?
For Laguna M.1, Q4_K_M (136.1 GB) offers the best balance of quality and VRAM usage. Q5_K_S (155.8 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XS at 68.3 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Laguna M.1 on a Mac?
Yes — Mac Studio (M3 Ultra, 96GB) and 20 other Macs can run Laguna M.1, the smallest at IQ3_XXS. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 136.1 GB of usable unified memory (RAM minus macOS overhead), from Mac Pro M2 Ultra (192 GB).
Can I run Laguna M.1 locally?
Yes — Laguna M.1 can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Laguna M.1?
At Q4_K_M, Laguna M.1 can reach ~59 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)
Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.
Example: NVIDIA B200 → 14.1 GB active ÷ (8000 × 0.65) = 2.72 ms, plus 70 layers × 0.055 ms = 3.85 ms, so 1000 ÷ 6.57 ms = ~152 tok/s
Estimated speed at Q4_K_M (136.1 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Laguna M.1?
At Q4_K_M, the download is about 135.48 GB. The full-precision BF16 version is 451.59 GB. The smallest option (IQ2_XS) is 67.74 GB.
Which GPUs can run Laguna M.1?
No single consumer GPU has enough VRAM to run Laguna M.1 at Q4_K_M (136.1 GB). Multi-GPU or professional hardware is required.
Which devices can run Laguna M.1?
4 devices with unified memory can run Laguna M.1 at Q4_K_M (136.1 GB), including Mac Pro M2 Ultra (192 GB), Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M2 Ultra (192 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.