Can I run Hy3 on an RTX 3090 (24GB)?
Computed with the open FitLLM engine — accurate per-layer KV-cache modeling, not a naive estimate. Updated 2026-07-10.
Memory breakdown (Q4_K_M, F16 KV, 33K context)
| Model weights | 170 GB |
| KV cache | 10.0 GB |
| Runtime overhead + reserve | 24.9 GB |
| Total used | 205 / 24 GB |
| Short by | 181 GB |
Max context that fits at Q4_K_M: does not fit.
Every quantization on the RTX 3090
| Weight quant | Weights | Fits (KV F16) | Used @32K |
|---|---|---|---|
| Q4_K_M | 170 GB | ❌ won't fit | 205 / 24.0 GB |
| Q5_K_M | 198 GB | ❌ won't fit | 237 / 24.0 GB |
| Q6_K | 228 GB | ❌ won't fit | 270 / 24.0 GB |
| Q8_0 | 296 GB | ❌ won't fit | 346 / 24.0 GB |
| FP16 | 557 GB | ❌ won't fit | 638 / 24.0 GB |
Lower weight quants free memory at some output-quality cost — Q4 is the common sweet spot; below that quality drops faster.
KV cache is F16 here (llama.cpp default). Drop it to Q8/Q4 (-ctk/-ctv) for more context.
Embed this verdict
Live badge for your README or model card — recomputed by the engine, never stale:
[](https://fitllm.run/can-i-run/hy3-on-rtx-3090)
Or from your terminal (exit 0/1 — works as a pre-download guard):
npx fitllm "Hy3" --gpu "RTX 3090"Why most VRAM calculators get this wrong
Hy3 is a Mixture-of-Experts: ~21B of 298.8B parameters are active per token, but all 298.8B must sit in memory. Naive calculators that size memory off active params (or KV off all layers) get this wrong.
What fits on the RTX 3090 instead
same GPU Models that fit on the RTX 3090: GLM-4.7-Flash, gpt-oss-20b, Qwen 3.6 27B, Gemma 4 e2b, Gemma 4 e4b, Gemma 4 12b, Gemma 4 26b A4B, Llama-3.2-3B-Instruct, Llama-3.1-8B-Instruct, Qwen3-0.6B, Qwen3-1.7B, Llama-3.2-1B-Instruct, Gemma-3-1B-it.
same model GPUs that run Hy3: —.
Reproduce it
Hy3 = 298.8B (21B active, MoE), 80 layers. The RTX 3090 has 24GB / 936GB/s. Same math, open source: fitllm-engine. GGUF bpw from llama.cpp.
All numbers are computed by the open-source fitllm-engine (MIT) from official model config.json values — reproduce or audit them yourself. Estimates; real usage varies with runtime (llama.cpp / MLX / Ollama), driver and display. Found a mismatch? Report it. · FitLLM home