Small models.
Built in public.

Lattice is a research project building compact conversational models! from a 42M parameter GPT trained from scratch to a 1.5B fine tune. Research demos, not production AI.

From scratch

Lattice Mini

42M params

A tiny language model built entirely from scratch. Custom tokenizer, pretraining, and instruction tuning. Expect wrong answers, that's the point at this scale.

Fine-tuned · Local

Lattice Spark

1.5B params

Qwen2.5-1.5B fine-tuned on Mac with MLX. Genuinely owns the Lattice identity — says "Lattice Systems" with no system prompt. Identity 7/8, factual 7/8, trained in 35 seconds for $0.

Benchmarks

Honest numbers across Mini, Pulse, and Pulse 2. The headline finding: a rank-16 QLoRA has no measurable effect on identity, but caused zero catastrophic forgetting (factual 100% everywhere).

Pulse 2 vs base Qwen3-8B QLoRA identity + factual eval — with & without system prompt
Benchmark chart comparing Lattice Pulse 2 (LoRA on Qwen3-8B) against base Qwen3-8B on identity and factual tasks, with and without the system prompt

All benchmarks & methodology →