mGPT 1.3B Uzbek
A 1.3B-parameter GPT-2-style model fine-tuned on Uzbek text for 50,000 steps on a single A100. Covers Uzbek, Russian, and English generation. It is a base model only — no instruction tuning.
If you need Uzbek language support at minimal VRAM cost, this is one of the very few options that exist at all — scarcity is doing some of the work here. It is a base model, so plan for prompt engineering or a fine-tuning layer before putting it in front of users. Community adoption is near zero, meaning bugs and edge cases are mostly undocumented. Hedge: useful as a research or prototype tool, but not production-ready out of the box.
›Why this rating
Auto-generated rating (Opus 4.7 judge, claude-opus-4-7). Overall 9.40/10. License (MIT) is explicit in the card and correctly flagged commercial-OK. Params (1.3B), context (2048), vendor, and GPT-2/GPT-3-style architecture all verify against the card. Description is honest and operator-voiced, correctly flagging it as a base model with no instruction tuning, and weaknesses appropriately call out low community traction and reasoning limits. Use case is sharp (Uzbek text generation), and the verdict properly hedges on production readiness. Brand fit is moderate — niche language model with limited audience — but the scarcity argument is legitimate and the row is honest about it.
Overview
A 1.3B-parameter GPT-2-style model fine-tuned on Uzbek text for 50,000 steps on a single A100. Covers Uzbek, Russian, and English generation. It is a base model only — no instruction tuning.
Strengths
- Genuine Uzbek-language coverage, rare at this weight class
- Also handles Russian and English in the same model
- MIT license — fully commercial-friendly
- Tiny footprint; runs on modest consumer hardware
Weaknesses
- Not instruction-tuned — requires prompting know-how or fine-tuning for chat/task use
- 2048-token context is tight by current standards
- 779 downloads and 13 likes signal very limited community testing
- 1.3B parameters will struggle with anything requiring multi-step reasoning
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 | 0.7 GB | 1 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 mGPT 1.3B Uzbek.
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 mGPT 1.3B Uzbek?
Can I use mGPT 1.3B Uzbek commercially?
What's the context length of mGPT 1.3B Uzbek?
Source: huggingface.co/ai-forever/mGPT-1.3B-uzbek
Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.
Related — keep moving
Verify mGPT 1.3B Uzbek runs on your specific hardware before committing money.