FA

90m Instruct

FalconGGUF by tiiuaeUpdated 2026-01-15
Download on Hugging Face
Parameters91M
Context256K
Layers24
Architecturefalcon-h1
Smallest quant0.0 GB
Quick answer

90m Instruct needs about 0.7 GB of memory at Q4_K_M (a 0.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 27 quants, from 0.0 GB (IQ2_M) to 0.2 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 0.4 GB.

Falcon‑H1 Tiny 90M Instruct – lightweight instruction‑tuned model

Falcon‑H1 Tiny 90M Instruct is a 90 million‑parameter causal decoder‑only model trained on English text. It supports instruction following and is available in multiple GGUF quantizations for llama.cpp and other inference engines.

Good at

  • Instruction following
  • Lightweight inference
  • Fast generation

Limits

  • No vision support
  • Requires 8 k context for estimates
  • License is Falcon‑LLM

Best for

  • Quick prototyping
  • Edge device inference
  • Low‑resource chatbots

Who it’s for. Developers needing a small, instruction‑tuned model for rapid experimentation or deployment on limited hardware. The model is suitable for projects that require fast inference with minimal memory footprint.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_M0.0 GB 0.6 GB
Fits Get file →
IQ3_XXS0.0 GB 0.6 GB
Fits Get file →
TQ1_00.0 GB 0.6 GB
Fits Get file →
TQ2_00.0 GB 0.6 GB
Fits Get file →
Q2_K0.0 GB 0.6 GB
Fits Get file →
IQ3_XS0.0 GB 0.6 GB
Fits Get file →
Q3_K_S0.0 GB 0.6 GB
Fits Get file →
IQ3_S0.0 GB 0.6 GB
Fits Get file →
IQ3_M0.0 GB 0.6 GB
Fits Get file →
Q3_K0.0 GB 0.7 GB
Fits Get file →
Q3_K_M0.0 GB 0.7 GB
Fits Get file →
Q3_K_L0.1 GB 0.7 GB
Fits Get file →
IQ4_XS0.1 GB 0.7 GB
Fits Get file →
Q4_00.1 GB 0.7 GB
Fits Get file →
Q4_K_S0.1 GB 0.7 GB
Fits Get file →
IQ4_NL0.1 GB 0.7 GB
Fits Get file →
Q4_K0.1 GB 0.7 GB
Fits Get file →
Q4_K_M0.1 GB 0.7 GB
Fits Get file →
Q4_10.1 GB 0.7 GB
Fits Get file →
Q5_00.1 GB 0.7 GB
Fits Get file →
Q5_K_S0.1 GB 0.7 GB
Fits Get file →
Q5_K0.1 GB 0.7 GB
Fits Get file →
Q5_K_M0.1 GB 0.7 GB
Fits Get file →
Q5_10.1 GB 0.7 GB
Fits Get file →
Q6_K0.1 GB 0.7 GB
Fits Get file →
Q8_00.1 GB 0.7 GB
Fits Get file →
BF160.2 GB 0.8 GB
Fits Get file →

For 8 GB VRAM, use Q4_K or Q5_K variants (~0.7 GB memory). With 4 GB VRAM, choose Q4_0 or Q4_K_S (~0.7 GB). For 2 GB VRAM, IQ3_S or IQ3_M (~0.6 GB) are recommended. File sizes range from 0.01 GB to 0.18 GB.

What GPU do you need for 90m Instruct?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is BF16 (0.2 GB file, about 0.8 GB in use).

Which 90m Instruct quant should you choose?

Q4_K_M is the usual starting point for 90m Instruct: it balances file size and quality well. Here it is a 0.1 GB file.

The Q8_0 file is 0.1 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is IQ2_M at 0.0 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturefalcon-h1
Parameters91M
Layers24
Trained context262,144 tokens
KV cache (fp16)12 KB per token

Source

Quantized bytiiuae
FamilyFalcon
Licenseother
Downloads (HF)5,242
Last updated2026-01-15

Frequently asked questions about 90m Instruct

How much VRAM does 90m Instruct need?

About 0.7 GB at Q4_K_M with an 8K context window. That is the 0.1 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run 90m Instruct on an 8 GB GPU?

Yes, with BF16: the 0.2 GB file needs about 0.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of 90m Instruct GGUF?

It depends on the quant: IQ2_M is 0.0 GB, IQ3_XXS is 0.0 GB, TQ1_0 is 0.0 GB, TQ2_0 is 0.0 GB, Q2_K is 0.0 GB, IQ3_XS is 0.0 GB and 21 more.

What context length does 90m Instruct support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.1 GB at 8K and 0.4 GB at 32K.

Where can I download 90m Instruct GGUF?

The files are on Hugging Face in the tiiuae/Falcon-H1-Tiny-90M-Instruct-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does 90m Instruct use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Falcon‑H1 Tiny 90M Instruct on 4 GB VRAM?

Yes, use a Q4_K or Q4_K_S quantization; memory usage is about 0.7 GB at 8 k context.

Does this model support image inputs?

No, the model is purely text‑based and has no vision capabilities.

What license applies to this model?

The model is released under the Falcon‑LLM License.

Similar models

Browse all →

More from tiiuae

All by tiiuae →

More Falcon models

All Falcon →

Other tiny models

All tiny models →