Maxwell V1 Flash — Hardware Requirements & GPU Compatibility
ChatFunctionsMaxwell V1 Flash is a 36.8B-parameter open language model from maxwell-model. It supports a context window of up to 524,288 tokens. At Q4_K_M it needs about 22.48 GB of VRAM — see which GPUs and Macs can run it below.
Decent on a 24 GB GPU
Runs on 24 GB (Q4_K_M · ~148 tok/s · 32K)Macs from 24 GB (Q3_K_M · ~51 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Maxwell V1 Flash
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run hf.co/AtomicChat/Maxwell-V1-Flash-GGUFQ4_K_M fits fully on the RTX 4090 · 0.9 GB free · 16K context
Download: Maxwell-V1-Flash-Q4_K_M.gguf22.09 GBAtomicChat
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX PRO 4500 (32 GB) (8 GB headroom) · Q4_K_M
- Generation speed
- ~138 tok/s
- generation speed
- Cost per 1M output tokens
- $0.69
- per 1M output tokens
How Much VRAM Does Maxwell V1 Flash Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 16.0 GB | 17.5 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 128K context · ~181 tok/s | |
| 18.3 GB | 19.8 GB | 3-bit medium — 3-bit medium quantization | Fits: 64K context · ~168 tok/s | |
| 20.2 GB | 21.6 GB | 4-bit · i-quant compact — Importance-weighted 4-bit, compact | Fits: 64K context · ~159 tok/s | |
| 22.5 GB | 23.9 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 32K context · ~148 tok/s | |
| 26.6 GB | 28 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Does not fit: needs 27.3 GB | |
| 30.8 GB | 32.2 GB | 6-bit, very good — 6-bit quantization, very good quality | Does not fit: needs 31.4 GB | |
| 37.2 GB | 38.6 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Does not fit: needs 37.9 GB | |
| 74.0 GB | 75.4 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 74.7 GB |
Selected Q4_K_M
4-bit medium quantization23.1GBat 16K context
- Weights
- 22.1 GB
- KV cache
- 0.8 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 0.9 GB free · ~148 tok/s
≥50 tok/s needs ≥245 GB/s · RTX 5090 ~205, RTX 3090 Ti ~148
Or .
Which Hardware Can Run Maxwell V1 Flash?
Q4_K_M · 22.5 GBMaxwell V1 Flash (Q4_K_M) requires 22.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 30+ GB is recommended. Using the full 524K context window can add up to 24.1 GB, bringing total usage to 46.5 GB. 7 GPUs can run it, including NVIDIA GeForce RTX 5090. 112 devices with unified memory can run Maxwell V1 Flash, including Mac Studio M5 Ultra (256 GB), Mac Studio M5 Ultra (512 GB), NVIDIA RTX Spark N1X Laptop (32 GB).
20 GB
Q4_K_M · best fit Q3_K_M (19.0 GB)
1 GPU
RX 7900 XT~139Simulate output on RX 7900 XT32–40 GB
fits · 8.9 GB free · up to 128K ctx
1 GPU · 3 Mac classes · 9 device families
RTX 5090~205Simulate output on RTX 5090M5 Max*36 GB~57Simulate output on Mac Studio M5 Max (36 GB)M632 GB~31Simulate output on iMac 24" M6 (32 GB)RTX Spark N1X~60Simulate output on NVIDIA RTX Spark N1X Laptop (32 GB)Orin~43Simulate output on NVIDIA Jetson AGX Orin 32GB
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.6 35B A3BAlibaba | 106.8 | 22 GB | ~151 | same size |
| MiniCPM V 4.7 35B A3BOpenBMB | — | 22.1 GB | ~169 | same size |
| Ornith 1.5 35B A3Bornith-ai | — | 22 GB | ~151 | same size |
| Agents A1InternScience | — | 21.5 GB | ~171 | same size |
| Intern S2 PreviewInternLM | — | 73.2 GB | —needs > 24 GB | same size |
| EXAONE 4.5 33BLGAI-EXAONE | — | 23.9 GB | ~33 | same size |
Specifications
- Publisher
- maxwell-model
- Parameters
- 36.8B
- Active parameters
- 4.6B per token
- Architecture
- MaxwellMoeForCausalLM
- Context Length
- 524,288 tokens
- Vocabulary Size
- 200,704
- Release Date
- 2026-10-13
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Maxwell V1 Flash need?
Maxwell V1 Flash requires 22.5 GB of VRAM at Q4_K_M, or 74.0 GB at BF16. Full 524K context adds up to 24.1 GB (46.5 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 36.8B × 4.8 bits ÷ 8 = 22.1 GB
KV Cache + Overhead ≈ 0.4 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 ≈ 24.4 GB (at full 524K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 4090 run Maxwell V1 Flash?
Yes, at Q4_K_M (22.5 GB) or lower. Higher quantizations like Q5_K_M (26.6 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
What's the best quantization for Maxwell V1 Flash?
For Maxwell V1 Flash, Q4_K_M (22.5 GB) offers the best balance of quality and VRAM usage. Q5_K_M (26.6 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 16.0 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Maxwell V1 Flash on a Mac?
Yes — Mac Mini M5 Pro (24 GB) and 115 other Macs can run Maxwell V1 Flash, the smallest at Q3_K_M. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 22.5 GB of usable unified memory (RAM minus macOS overhead), from Mac Studio M1 Max (32 GB).
Can I run Maxwell V1 Flash locally?
Yes — Maxwell V1 Flash can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Maxwell V1 Flash?
At Q4_K_M, Maxwell V1 Flash can reach ~327 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~148 tok/s. 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 → 2.8 GB active ÷ (8000 × 0.65) = 0.537 ms, plus 45 layers × 0.055 ms = 2.475 ms, so 1000 ÷ 3.012 ms = ~332 tok/s
Estimated speed at Q4_K_M (22.5 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Maxwell V1 Flash?
At Q4_K_M, the download is about 22.09 GB. The full-precision BF16 version is 73.62 GB. The smallest option (Q2_K) is 15.64 GB.
Which GPUs can run Maxwell V1 Flash?
7 consumer GPUs can run Maxwell V1 Flash at Q4_K_M. Top options include AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090.
Which devices can run Maxwell V1 Flash?
112 devices with unified memory can run Maxwell V1 Flash at Q4_K_M (22.5 GB), including AMD Ryzen 7 8845HS (Hawk Point) Laptop, AMD Ryzen AI 7 350 (Krackan Point) Laptop, AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10. Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.