Text model · Ornith

OR Ornith 1.0 35B: RAM and VRAM requirements

Ornith 1.0 35B needs about 23.2 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~21 GB to download; KV cache and overhead add the rest), or about 39.2 GB at Q8_0. The lightest hardware that runs it is Nvidia GeForce RTX 5090 (32GB).

Ornith family · 35B params (Mixture-of-Experts: activates only 3B of 35B params per token, so generation is faster than the total size suggests) · released Jun 2026.

Shopping for hardware? See what runs Ornith 1.0 35B →

License MIT · Commercial OK ↓ 2.8M/mo ♥ 462 on HuggingFace

Q4_K_M GGUF

21 GB

Q8_0 GGUF

37 GB

Memory @ Q4 (4k)

~23.2 GB

Context

256 k

Q4_K_M and Q8_0 are GGUF download (file) sizes. Memory @ Q4 includes the KV cache and overhead, so it runs larger than the file on disk.

Will it run on your device?

Ornith 1.0 35B runs on 9 of 40 tracked devices at Q4_K_M.

9 run well 0 tight fit 31 too small

Fit check Q4_K_M

Your device

Apple M4 (16GB)

No, not enough memory

needs 23.2 GB usable 10.5 GB

See the full breakdown

Memory breakdown

Weights (Q4_K_M)21 GB

KV cache (4k)1.4 GB

Overhead0.8 GB

Total23.2 GB

How context length changes it

4k context ~23.2 GB 32k context ~33.2 GB 128k context ~67.2 GB

Longer context grows the KV cache, which for Ornith 1.0 35B is sized by its 3B active params, not the full 35B. It needs ~23.2 GB at 4k and ~67.2 GB at 128k.

Quantization sizes

GGUF quantson disk

Quantization	Size on disk
Q2_K	14.7 GB est
Q3_K_M	17.1 GB est
Q4_K_M (default)	21 GB
Q5_K_M	24.9 GB est
Q6_K	28.7 GB est
Q8_0	37 GB
FP16	69 GB

Lower quant = smaller and faster, slightly lower quality. Q4_K_M is the common default.

Run it

Ollama

$ ollama run ornith:35b

llama.cpp

$ llama-cli -hf deepreinforce-ai/Ornith-1.0-35B-GGUF:Q4_K_M

LM Studio

$ lms get deepreinforce-ai/Ornith-1.0-35B-GGUF

Which devices can run Ornith 1.0 35B?

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

Android

NVIDIA GPUs

AMD GPUs

Same job, different size

Pick by what fits your memory: step down to free up VRAM, or step up if you have headroom.

Next size down Qwen-AgentWorld 35B-A3B 34.7B Next size up Seed-OSS 36B Instruct 36B

Similar models

Head-to-head

Ornith 1.0 35B vs Ornith 1.0 9B

FAQ

How much VRAM or RAM does Ornith 1.0 35B need?

At Q4_K_M, Ornith 1.0 35B needs about 23.2 GB (weights ~21 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~39.2 GB.

What is the Q4_K_M GGUF file size of Ornith 1.0 35B?

The Q4_K_M GGUF file is about 21 GB to download, and the Q8_0 GGUF is about 37 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~23.2 GB of memory at Q4_K_M.

Can Ornith 1.0 35B run on a laptop?

Ornith 1.0 35B is large; you need a 24 GB+ GPU or a 32-48 GB Mac at Q4_K_M.

Is Ornith 1.0 35B cheaper to run because it is a MoE model?

It is faster, not lighter. Ornith 1.0 35B activates only 3B of 35B params per token (so it runs quickly), but all experts must stay in memory, so it still needs memory for the full 35B.

Can I use Ornith 1.0 35B commercially?

Yes. Ornith 1.0 35B is licensed MIT, which permits commercial use.

Understand the numbers

Short guides to the ideas behind Ornith 1.0 35B's memory and quant figures.

DeepReinforce's larger coding-agent MoE: 35B total, 3B active (built on the Qwen3.6 35B-A3B base), 256K context, MIT. Fast on unified memory thanks to the 3B active params.

Sources

Last validated 2026-08-03. Memory figures are estimates. See methodology.