// model_catalog

29 models that run on Little Brother

Every model is quantized for Apple Silicon, tested by us, and runs fully on-device. Grouped by where it runs, then by family. Memory figures are on-device peaks — weights are download size estimates.

iPhone & iPad

18 models

Models that fit within iOS's ~5 GB per-app memory budget. “Tight” models still run, but sit close to the limit on some devices.

Qwen 6

Qwen2.5 0.5B Instruct (4-bit)

Ultra-light chat and quick PDF Q&A; long answers cut off

RAM peak: 1.5–2.3 GB
Weights: 0.27 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Alibaba (Qwen) Apache 2.0

Qwen3 0.6B (4-bit)

Fast reasoning, PDF Q&A, and doc lookup at a tiny size

RAM peak: 1.5–2.3 GB
Weights: 0.33 GB
Context: 40,960 tok
Min Mac RAM: 8 GB

Reasoning

Alibaba (Qwen) Apache 2.0

Qwen2.5 1.5B Instruct (4-bit)

Solid everyday chat and PDF work; doc lookup can miss specifics

RAM peak: 2.0–2.8 GB
Weights: 0.82 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Alibaba (Qwen) Apache 2.0

Qwen3 1.7B (4-bit)

Reasoning plus accurate doc lookup in a light model

RAM peak: 2.1–2.9 GB
Weights: 0.92 GB
Context: 40,960 tok
Min Mac RAM: 8 GB

Reasoning

Alibaba (Qwen) Apache 2.0

Qwen3.5 2B (4-bit)

Current-gen Qwen chat with reasoning; forgot cross-chat facts in tests

RAM peak: 2.8–3.6 GB
Weights: 1.63 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

Reasoning

Alibaba (Qwen) Apache 2.0

Qwen3 4B Instruct 2507 (4-bit)

Top all-rounder — reasoning, web search, PDF tools; passed every test

RAM peak: 3.3–4.1 GB
Weights: 2.12 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

Reasoning Tools Web search Documents

Alibaba (Qwen) Apache 2.0

LFM (Liquid AI) 1

LFM2.5 1.2B Instruct (4-bit)

Fast, dependable chat and PDF/doc Q&A (Liquid AI)

RAM peak: 1.8–2.6 GB
Weights: 0.62 GB
Context: 128,000 tok
Min Mac RAM: 8 GB

Liquid AI LFM Open License v1.0

Llama 2

Llama 3.2 1B Instruct (4-bit)

Light general chat and PDF summaries; weak doc lookup in tests

RAM peak: 1.9–2.7 GB
Weights: 0.66 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

Meta Llama 3.2 Community License

Llama 3.2 3B Instruct (4-bit)

Reliable general chat and doc Q&A; PDF summaries occasionally thin

RAM peak: 2.9–3.7 GB
Weights: 1.7 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

Tools

Meta Llama 3.2 Community License

Gemma 3

Gemma 3 1B Instruct QAT (4-bit)

Quality chat at a tiny footprint (QAT); mixed doc lookup in tests

RAM peak: 1.9–2.7 GB
Weights: 0.72 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Google DeepMind Gemma Terms of Use

Gemma 3n E2B Instruct (4-bit)

Efficient chat with strong doc lookup; long answers sometimes cut off

RAM peak: 3.6–4.4 GB
Weights: 2.37 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Google DeepMind Gemma Terms of Use

Gemma 3 4B Instruct QAT (4-bit)

Higher-quality Gemma chat and PDF summaries (QAT-tuned)

RAM peak: 4.0–4.8 GB
Weights: 2.83 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

iOS: tight

Google DeepMind Gemma Terms of Use

SmolLM 1

SmolLM3 3B (4-bit)

Reasoning, PDFs, and doc lookup — passed every on-device test

RAM peak: 2.9–3.7 GB
Weights: 1.67 GB
Context: 65,536 tok
Min Mac RAM: 8 GB

Reasoning Tools Documents

Hugging Face Apache 2.0

Phi 2

Phi-3.5 Mini Instruct (4-bit)

Default model — balanced chat and PDF Q&A; long answers cut off in tests

RAM peak: 3.2–4.0 GB
Weights: 2.0 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

Microsoft MIT

Phi-4 Mini Instruct (4-bit)

Short chat and simple doc questions; unstable on iOS and long PDF tasks

RAM peak: 3.3–4.1 GB
Weights: 2.16 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

iOS: tight Tools Documents

Microsoft MIT

Nemotron 2

Nemotron 3 Nano 4B (4-bit)

Reasoning, web search, and doc Q&A — near-perfect in tests

RAM peak: 3.2–4.0 GB
Weights: 2.24 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License

Llama Nemotron Nano 4B v1.1 (4-bit)

Reasoning and math; skipped web search and failed PDF summaries in tests

RAM peak: 3.5–4.3 GB
Weights: 2.54 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

iOS: tight Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License + Llama 3.1 Community License

Ministral 1

Ministral 3 3B Instruct (4-bit)

Compact instruction-following chat and doc Q&A

RAM peak: 3.8–4.6 GB
Weights: 2.59 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

iOS: tight Tools Documents

Mistral AI Apache 2.0

Mac

29 models

Every catalog model runs on a Mac with enough memory — the card shows the minimum RAM each one needs.

Qwen 9

Qwen2.5 0.5B Instruct (4-bit)

Ultra-light chat and quick PDF Q&A; long answers cut off

RAM peak: 1.5–2.3 GB
Weights: 0.27 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Alibaba (Qwen) Apache 2.0

Qwen3 0.6B (4-bit)

Fast reasoning, PDF Q&A, and doc lookup at a tiny size

RAM peak: 1.5–2.3 GB
Weights: 0.33 GB
Context: 40,960 tok
Min Mac RAM: 8 GB

Reasoning

Alibaba (Qwen) Apache 2.0

Qwen2.5 1.5B Instruct (4-bit)

Solid everyday chat and PDF work; doc lookup can miss specifics

RAM peak: 2.0–2.8 GB
Weights: 0.82 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Alibaba (Qwen) Apache 2.0

Qwen3 1.7B (4-bit)

Reasoning plus accurate doc lookup in a light model

RAM peak: 2.1–2.9 GB
Weights: 0.92 GB
Context: 40,960 tok
Min Mac RAM: 8 GB

Reasoning

Alibaba (Qwen) Apache 2.0

Qwen3.5 2B (4-bit)

Current-gen Qwen chat with reasoning; forgot cross-chat facts in tests

RAM peak: 2.8–3.6 GB
Weights: 1.63 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

Reasoning

Alibaba (Qwen) Apache 2.0

Qwen3 4B Instruct 2507 (4-bit)

Top all-rounder — reasoning, web search, PDF tools; passed every test

RAM peak: 3.3–4.1 GB
Weights: 2.12 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

Reasoning Tools Web search Documents

Alibaba (Qwen) Apache 2.0

Qwen3.5 4B (4-bit)

Strong reasoning, web search, and PDF tools — Mac only (OOM on iOS)

RAM peak: 4.1–4.9 GB
Weights: 2.85 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

Mac only Reasoning Tools Web search Documents

Alibaba (Qwen) Apache 2.0

Qwen2.5 7B Instruct (4-bit)

High-accuracy PDF tools, web search, and doc lookup (needs 16 GB Mac)

RAM peak: 5.2–6.0 GB
Weights: 4.0 GB
Context: 32,768 tok
Min Mac RAM: 16 GB

Mac only Tools Web search Documents

Alibaba (Qwen) Apache 2.0

Qwen3 8B (4-bit)

Strongest reasoning in class — passed every on-device test

RAM peak: 5.5–6.3 GB
Weights: 4.31 GB
Context: 40,960 tok
Min Mac RAM: 16 GB

Mac only Reasoning Tools Web search Documents

Alibaba (Qwen) Apache 2.0

LFM (Liquid AI) 2

LFM2.5 1.2B Instruct (4-bit)

Fast, dependable chat and PDF/doc Q&A (Liquid AI)

RAM peak: 1.8–2.6 GB
Weights: 0.62 GB
Context: 128,000 tok
Min Mac RAM: 8 GB

Liquid AI LFM Open License v1.0

LFM2 8B A1B (3-bit)

Fast mixture-of-experts chat and PDFs on Mac; doc lookup missed specifics

RAM peak: 5.1–5.9 GB
Weights: 3.89 GB
Context: 128,000 tok
Min Mac RAM: 16 GB

Mac only Tools Documents

Liquid AI LFM Open License v1.0

Llama 3

Llama 3.2 1B Instruct (4-bit)

Light general chat and PDF summaries; weak doc lookup in tests

RAM peak: 1.9–2.7 GB
Weights: 0.66 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

Meta Llama 3.2 Community License

Llama 3.2 3B Instruct (4-bit)

Reliable general chat and doc Q&A; PDF summaries occasionally thin

RAM peak: 2.9–3.7 GB
Weights: 1.7 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

Tools

Meta Llama 3.2 Community License

Meta Llama 3.1 8B Instruct (4-bit)

Dependable general chat and doc Q&A on Mac

RAM peak: 5.4–6.2 GB
Weights: 4.22 GB
Context: 131,072 tok
Min Mac RAM: 16 GB

Mac only Tools Documents

Meta Llama 3.1 Community License

Gemma 3

Gemma 3 1B Instruct QAT (4-bit)

Quality chat at a tiny footprint (QAT); mixed doc lookup in tests

RAM peak: 1.9–2.7 GB
Weights: 0.72 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Google DeepMind Gemma Terms of Use

Gemma 3n E2B Instruct (4-bit)

Efficient chat with strong doc lookup; long answers sometimes cut off

RAM peak: 3.6–4.4 GB
Weights: 2.37 GB
Context: 32,768 tok
Min Mac RAM: 8 GB

Google DeepMind Gemma Terms of Use

Gemma 3 4B Instruct QAT (4-bit)

Higher-quality Gemma chat and PDF summaries (QAT-tuned)

RAM peak: 4.0–4.8 GB
Weights: 2.83 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

iOS: tight

Google DeepMind Gemma Terms of Use

SmolLM 1

SmolLM3 3B (4-bit)

Reasoning, PDFs, and doc lookup — passed every on-device test

RAM peak: 2.9–3.7 GB
Weights: 1.67 GB
Context: 65,536 tok
Min Mac RAM: 8 GB

Reasoning Tools Documents

Hugging Face Apache 2.0

Phi 3

Phi-3.5 Mini Instruct (4-bit)

Default model — balanced chat and PDF Q&A; long answers cut off in tests

RAM peak: 3.2–4.0 GB
Weights: 2.0 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

Microsoft MIT

Phi-4 Mini Instruct (4-bit)

Short chat and simple doc questions; unstable on iOS and long PDF tasks

RAM peak: 3.3–4.1 GB
Weights: 2.16 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

iOS: tight Tools Documents

Microsoft MIT

Phi-4 (3-bit)

High-accuracy reasoning and math on Mac

RAM peak: 7.0–7.8 GB
Weights: 5.98 GB
Context: 16,384 tok
Min Mac RAM: 16 GB

Mac only

Microsoft MIT

Nemotron 7

Nemotron 3 Nano 4B (4-bit)

Reasoning, web search, and doc Q&A — near-perfect in tests

RAM peak: 3.2–4.0 GB
Weights: 2.24 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License

Llama Nemotron Nano 4B v1.1 (4-bit)

Reasoning and math; skipped web search and failed PDF summaries in tests

RAM peak: 3.5–4.3 GB
Weights: 2.54 GB
Context: 131,072 tok
Min Mac RAM: 8 GB

iOS: tight Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License + Llama 3.1 Community License

Llama Nemotron 8B UltraLong 1M (4-bit)

Long-context reasoning, web search, and doc Q&A on Mac

RAM peak: 5.5–6.3 GB
Weights: 4.52 GB
Context: 1,073,152 tok
Min Mac RAM: 16 GB

Mac only Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License + Llama 3.1 Community License

Nemotron Nano 9B v2 (4-bit)

Reasoning, web search, and docs — passed every on-device test

RAM peak: 6.0–6.8 GB
Weights: 5.0 GB
Context: 131,072 tok
Min Mac RAM: 16 GB

Mac only Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License

Nemotron 3 Nano 30B A3B (4-bit)

MoE flagship — ~3B active/token, strong reasoning on Mac

RAM peak: 18.8–19.6 GB
Weights: 17.8 GB
Context: 262,144 tok
Min Mac RAM: 32 GB

Mac only Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License

Llama Nemotron 70B Instruct (4-bit)

Top Llama Nemotron instruct quality on high-RAM Macs

RAM peak: 40.7–41.5 GB
Weights: 39.7 GB
Context: 131,072 tok
Min Mac RAM: 64 GB

Mac only Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License + Llama 3.1 Community License

Nemotron 3 Super 120B A12B (4-bit)

Largest Nemotron 3 — ~12B active/token for M3 Ultra class Macs

RAM peak: 121.7–122.5 GB
Weights: 120.7 GB
Context: 262,144 tok
Min Mac RAM: 128 GB

Mac only Reasoning Tools Web search Documents

NVIDIA NVIDIA Open Model License

Ministral 1

Ministral 3 3B Instruct (4-bit)

Compact instruction-following chat and doc Q&A

RAM peak: 3.8–4.6 GB
Weights: 2.59 GB
Context: 262,144 tok
Min Mac RAM: 8 GB

iOS: tight Tools Documents

Mistral AI Apache 2.0

// sovereign_vault_protocol

Enter the Sovereign Vault

Little Brother is coming soon. No account required. Nothing leaves your device.

Coming Soon