Gemma 4 26B A4B IT Uncensored Heretic — Hardware Requirements & GPU Compatibility
Vision
Gemma 4 26B A4B IT Uncensored Heretic is a 25.8B-parameter open language model from llmfan46 in the Gemma 4 family. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 16.13 GB of VRAM — see which GPUs and Macs can run it below.
2.1K downloads
12 likesSpecifications
- Publisher
- llmfan46
- Family
- Gemma 4
- Parameters
- 25.8B
- Architecture
- Gemma4ForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 262,144
- Release Date
- 2026-04-07
- License
- Apache 2.0
Get Started
How Much VRAM Does Gemma 4 26B A4B IT Uncensored Heretic Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_Kest. | 3.40 | 11.6 GB | 55.6 GB | 10.97 GB | 2-bit quantization with K-quant improvements |
| Q3_K_M | 3.90 | 13.2 GB | 57.2 GB | 12.58 GB | 3-bit medium quantization |
| Q3_K_L | 4.10 | 13.9 GB | 57.8 GB | 13.23 GB | 3-bit large quantization |
| Q4_K_S | 4.50 | 15.2 GB | 59.1 GB | 14.52 GB | 4-bit small quantization |
| Q4_K_M | 4.80 | 16.1 GB | 60.1 GB | 15.48 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_S | 5.50 | 18.4 GB | 62.3 GB | 17.74 GB | 5-bit small quantization |
| Q5_K_M | 5.70 | 19.0 GB | 63.0 GB | 18.39 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 21.9 GB | 65.9 GB | 21.29 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 26.4 GB | 70.4 GB | 25.81 GB | 8-bit quantization, near-lossless |
| BF16 | 16.00 | 52.3 GB | 96.2 GB | 51.61 GB | Brain floating point 16 — preferred for training |
est.= calculated VRAM estimate; no published GGUF file found for that quantization yet. Other rows are verified against real community uploads.
Which GPUs Can Run Gemma 4 26B A4B IT Uncensored Heretic?
Q4_K_M · 16.1 GBGemma 4 26B A4B IT Uncensored Heretic (Q4_K_M) requires 16.1 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 21+ GB is recommended. Using the full 262K context window can add up to 44.0 GB, bringing total usage to 60.1 GB. 6 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroom
Which Devices Can Run Gemma 4 26B A4B IT Uncensored Heretic?
Q4_K_M · 16.1 GB21 devices with unified memory can run Gemma 4 26B A4B IT Uncensored Heretic, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Mini M4 Pro (24 GB).
Runs great
— Plenty of headroomWhere to Download Gemma 4 26B A4B IT Uncensored Heretic
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Related Models
Frequently Asked Questions
- How much VRAM does Gemma 4 26B A4B IT Uncensored Heretic need?
Gemma 4 26B A4B IT Uncensored Heretic requires 16.1 GB of VRAM at Q4_K_M, or 52.3 GB at BF16. Full 262K context adds up to 44.0 GB (60.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 25.8B × 4.8 bits ÷ 8 = 15.5 GB
KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 44.6 GB (at full 262K context)
VRAM usage by quantization
Q4_K_M16.1 GBQ4_K_M + full context60.1 GB- Can NVIDIA GeForce RTX 4090 run Gemma 4 26B A4B IT Uncensored Heretic?
Yes, at Q6_K (21.9 GB) or lower. Higher quantizations like Q8_0 (26.4 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
- What's the best quantization for Gemma 4 26B A4B IT Uncensored Heretic?
For Gemma 4 26B A4B IT Uncensored Heretic, Q4_K_M (16.1 GB) offers the best balance of quality and VRAM usage. Q5_K_S (18.4 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 11.6 GB.
VRAM requirement by quantization
Q2_K11.6 GBQ3_K_L13.9 GBQ4_K_M ★16.1 GBQ5_K_S18.4 GBQ6_K21.9 GBBF1652.3 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Gemma 4 26B A4B IT Uncensored Heretic on a Mac?
Gemma 4 26B A4B IT Uncensored Heretic requires at least 11.6 GB at Q2_K, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Gemma 4 26B A4B IT Uncensored Heretic locally?
Yes — Gemma 4 26B A4B IT Uncensored Heretic can run locally on consumer hardware. At Q4_K_M quantization it needs 16.1 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Gemma 4 26B A4B IT Uncensored Heretic?
At Q4_K_M, Gemma 4 26B A4B IT Uncensored Heretic can reach ~181 tok/s on AMD Instinct MI300X. On NVIDIA GeForce RTX 4090: ~41 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 16.1 × 0.55 = ~181 tok/s
Estimated speed at Q4_K_M (16.1 GB)
~181 tok/s~41 tok/s~135 tok/s~112 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Gemma 4 26B A4B IT Uncensored Heretic?
At Q4_K_M, the download is about 15.48 GB. The full-precision BF16 version is 51.61 GB. The smallest option (Q2_K) is 10.97 GB.
- Which GPUs can run Gemma 4 26B A4B IT Uncensored Heretic?
6 consumer GPUs can run Gemma 4 26B A4B IT Uncensored Heretic at Q4_K_M (16.1 GB). Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XT, AMD Radeon RX 7900 XTX. 1 GPU have plenty of headroom for comfortable inference.
- Which devices can run Gemma 4 26B A4B IT Uncensored Heretic?
21 devices with unified memory can run Gemma 4 26B A4B IT Uncensored Heretic at Q4_K_M (16.1 GB), including Mac Mini M4 (32 GB), Mac Mini M4 Pro (24 GB), Mac Mini M4 Pro (48 GB), Mac Pro M2 Ultra (192 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.