Gemma 4 Turkish 26B (4B active)
Gemma 4 26B MoE (4B active params) pruned and Turkish-tuned. The largest Turkish-tuned open-weight model on HF as of May 2026. MoE architecture means it loads 26B of weights but runs at 4B-active speed.
Overview
Gemma 4 26B MoE (4B active params) pruned and Turkish-tuned. The largest Turkish-tuned open-weight model on HF as of May 2026. MoE architecture means it loads 26B of weights but runs at 4B-active speed.
Strengths
- Largest Turkish open-weight model available (26B total)
- MoE architecture — 4B active means inference speed closer to a 4-7B dense model
- 128K context — usable for long-document Turkish tasks
Weaknesses
- Requires significant VRAM (~14GB even at Q4) to load all experts
- Community release; limited public benchmarks vs the YTU and Trendyol mainline models
- MoE quantization is noisier than dense at Q4 — prefer Q5/Q6 if VRAM allows
Quantization variants
Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.
| Quantization | File size | VRAM required |
|---|---|---|
| Q4_K_M | 14.3 GB | 19 GB |
Get the model
HuggingFace
Original weights
Source repository — direct quantization required.
Hardware that runs this
Cards with enough VRAM for at least one quantization of Gemma 4 Turkish 26B (4B active).
Models worth comparing
Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.
Frequently asked
What's the minimum VRAM to run Gemma 4 Turkish 26B (4B active)?
Can I use Gemma 4 Turkish 26B (4B active) commercially?
What's the context length of Gemma 4 Turkish 26B (4B active)?
Source: huggingface.co/esokullu/gemma4-turkish-26b-a4b-pruned-gguf
Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.
Related — keep moving
Verify Gemma 4 Turkish 26B (4B active) runs on your specific hardware before committing money.