Qwen3.8 Flash Next REAM 60Pct — Hardware Requirements & GPU Compatibility
Chat
Qwen3.8 Flash Next REAM 60Pct is a 128.8B-parameter open language model from Akicou in the Qwen 3.8 family. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 77.66 GB of VRAM — see which GPUs and Macs can run it below.
Qwen3.8 Flash Next REAM 60Pct runs on a 96 GB+ Mac or a workstation GPU
Best local pick → Mac Studio (M3 Ultra, 96GB) · Q4_K_M · 78.3 GB · ~14 tok/s
Macs from 64 GB (Q2_K · ~18 tok/s)
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Qwen3.8 Flash Next REAM 60Pct
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
ollama run hf.co/Akicou/Qwen3.8-Flash-Next-REAM-60Pct-GGUF:Q4_K_MQ4_K_M needs offload on the RTX 4090 · 54.3 GB short · 16K context
Download: qwen4-ream-q4_k_m.gguf86.96 GBAkicou
Other ways
each opens step-by-stepRun in cloud
Fits on RTX PRO 6000 (96 GB) (17 GB headroom) · Q4_K_M
- Generation speed
- ~30 tok/s
- generation speed
- Cost per 1M output tokens
- $11.47
- per 1M output tokens
How Much VRAM Does Qwen3.8 Flash Next REAM 60Pct Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 55.1 GB | 56.4 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Does not fit: needs 55.7 GB | |
| 63.2 GB | 64.4 GB | 3-bit medium — 3-bit medium quantization | Does not fit: needs 63.8 GB | |
| 72.8 GB | 74.1 GB | 4-bit small — 4-bit small quantization | Does not fit: needs 73.4 GB | |
| 77.7 GB | 78.9 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Does not fit: needs 78.3 GB | |
| 92.2 GB | 93.4 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Does not fit: needs 92.7 GB | |
| 106.6 GB | 107.9 GB | 6-bit, very good — 6-bit quantization, very good quality | Does not fit: needs 107.2 GB | |
| 129.2 GB | 130.4 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Does not fit: needs 129.8 GB | |
| 258.0 GB | 259.2 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 258.6 GB |
Selected Q4_K_M
4-bit medium quantization78.3GBat 16K context
- Weights
- 77.3 GB
- KV cache
- 0.7 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 54.2 GB more
Which Hardware Can Run Qwen3.8 Flash Next REAM 60Pct?
Q4_K_M · 77.7 GBQwen3.8 Flash Next REAM 60Pct (Q4_K_M) requires 77.7 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 101+ GB is recommended. Using the full 262K context window can add up to 10.7 GB, bringing total usage to 88.3 GB. No consumer GPU has enough memory. 30 devices with unified memory can run Qwen3.8 Flash Next REAM 60Pct, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio (M3 Ultra, 96GB).
64–80 GB
Q4_K_M · best fit Q3_K_M (63.8 GB)
1 device family
Orin~5Simulate output on NVIDIA Jetson AGX Orin 64GB96 GB
fits · 17.8 GB free · up to 256K ctx
2 Mac classes
M3 Ultra*96 GB~14Simulate output on Mac Studio (M3 Ultra, 96GB)M3 Max*96 GB~6Simulate output on MacBook Pro 14" M3 Max (96 GB)
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to Qwen3.8 Flash Next REAM 60Pct
Qwen 3.8 family · 12 models →Similar models
5| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.5 122B A10BAlibaba | 109.4 | 75.5 GB | —needs > 24 GB | same size |
| Mistral Small 4 119B 2603Mistral AI | 96.4 | 81.6 GB | —needs > 24 GB | same size |
| NVIDIA Nemotron 3.5 Super VL 120B A12B BF16NVIDIA | — | 87.7 GB | —needs > 24 GB | same size |
| Ling 3.0 Flash VLInclusion AI | — | 76 GB | —needs > 24 GB | same size |
| XORTRON.CriminalComputing.LARGE.2026.3darkc0de | — | 79.8 GB | —needs > 24 GB | same size |
Specifications
- Publisher
- Akicou
- Family
- Qwen 3.8
- Parameters
- 128.8B
- Active parameters
- 58.5B per token
- Architecture
- Qwen4ExpForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-08-29
Frequently Asked Questions
How much VRAM does Qwen3.8 Flash Next REAM 60Pct need?
Qwen3.8 Flash Next REAM 60Pct requires 77.7 GB of VRAM at Q4_K_M, or 258.0 GB at BF16. Full 262K context adds up to 10.7 GB (88.3 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 128.8B × 4.8 bits ÷ 8 = 77.3 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 ≈ 11 GB (at full 262K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Qwen3.8 Flash Next REAM 60Pct?
No — Qwen3.8 Flash Next REAM 60Pct requires at least 55.1 GB at Q2_K, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
What's the best quantization for Qwen3.8 Flash Next REAM 60Pct?
For Qwen3.8 Flash Next REAM 60Pct, Q4_K_M (77.7 GB) offers the best balance of quality and VRAM usage. Q5_K_M (92.2 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 55.1 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Qwen3.8 Flash Next REAM 60Pct on a Mac?
Yes — Mac Pro M2 Ultra (64 GB) and 41 other Macs can run Qwen3.8 Flash Next REAM 60Pct, the smallest at Q2_K. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 77.7 GB of usable unified memory (RAM minus macOS overhead), from Mac Studio (M3 Ultra, 96GB).
Can I run Qwen3.8 Flash Next REAM 60Pct locally?
Yes — Qwen3.8 Flash Next REAM 60Pct can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Qwen3.8 Flash Next REAM 60Pct?
At Q4_K_M, Qwen3.8 Flash Next REAM 60Pct 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 → 35.3 GB active ÷ (8000 × 0.65) = 6.78 ms, plus 48 layers × 0.055 ms = 2.64 ms, so 1000 ÷ 9.42 ms = ~106 tok/s
Estimated speed at Q4_K_M (77.7 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Qwen3.8 Flash Next REAM 60Pct?
At Q4_K_M, the download is about 77.27 GB. The full-precision BF16 version is 257.58 GB. The smallest option (Q2_K) is 54.74 GB.
Which GPUs can run Qwen3.8 Flash Next REAM 60Pct?
No single consumer GPU has enough VRAM to run Qwen3.8 Flash Next REAM 60Pct at Q4_K_M (77.7 GB). Multi-GPU or professional hardware is required.
Which devices can run Qwen3.8 Flash Next REAM 60Pct?
30 devices with unified memory can run Qwen3.8 Flash Next REAM 60Pct at Q4_K_M (77.7 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.