Can I run Gemma 4 26b A4B on an RTX 3090 (24GB)?
Computed with the open FitLLM engine — accurate per-layer KV-cache modeling, not a naive estimate. Updated 2026-07-16.
Memory breakdown (Q4_K_M, F16 KV, 33K context)
| Model weights | 14.5 GB |
| KV cache | 0.8 GB |
| Runtime overhead + reserve | 4.8 GB |
| Total used | 20.2 / 24 GB |
| Free | 3.8 GB |
Max context that fits at Q4_K_M: ~83K tokens · with Q8 KV cache → ~108K tokens.
Every quantization on the RTX 3090
| Weight quant | Weights | Fits (KV F16) | Used @32K |
|---|---|---|---|
| Q4_K_M | 14.5 GB | ✅ up to 83K ctx | 20.2 / 24.0 GB |
| Q5_K_M | 16.9 GB | ⚠️ up to 31K ctx | 22.9 / 24.0 GB |
| Q6_K | 19.5 GB | ❌ won't fit | 25.7 / 24.0 GB |
| Q8_0 | 25.2 GB | ❌ won't fit | 32.2 / 24.0 GB |
| FP16 | 47.5 GB | ❌ won't fit | 57.1 / 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/gemma-4-26b-a4b-on-rtx-3090)
Or from your terminal (exit 0/1 — works as a pre-download guard):
npx fitllm "Gemma 4 26b A4B" --gpu "RTX 3090"Why most VRAM calculators get this wrong
Gemma 4 26b A4B interleaves sliding-window (local) and global attention 5:1. The local layers cap their KV cache at the 1024-token window, and the global layers use a different head shape (head_dim 512 vs 256). A naive "all layers × full context × one head_dim" formula over-counts KV cache by several times.
Other options
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, MiniCPM5-1B, Qwen3-0.6B, Qwen3-1.7B, Llama-3.2-1B-Instruct, Gemma-3-1B-it.
same model GPUs that run Gemma 4 26b A4B: RTX 5090 (32GB), RTX 4090 (24GB), RTX 3090 (24GB), RTX 3090 Ti (24GB), RTX 6000 Ada (48GB), RTX PRO 6000 Blackwell (96GB), RX 7900 XTX (24GB), RX 7900 XT (20GB), Radeon PRO W7900 (48GB), 2× RTX 3090 (48GB), 2× RTX 4090 (48GB), 4× RTX 3090 (96GB), A100 40GB (40GB), A100 80GB (80GB), H100 80GB (80GB), H200 141GB (141GB), B200 (180GB).
Reproduce it
Gemma 4 26b A4B = 25.5B (4B active, MoE), 30 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