BaseRT
6.4x faster than llama.cpp, 3.9x faster than MLX
Details
- External ID
- 1199683
- Source
- PH
- Company
- —
- Product
- BaseRT
- Website domain
- producthunt.com
- Launched
- July 19, 2026
- Cohort
- —
- Upvotes
- 222
- Upvotes percentile
- 0.3922018348623853
- Tags
- Open Source, Artificial Intelligence, Apple
- Fetched at
- Sept. 7, 2026, 1:22 a.m.
- Updated at
- Sept. 7, 2026, 1:22 a.m.
Description
BaseRT is the fastest LLM runtime on Apple Silicon. Install it with one command and run local models on your own device.
Enrichment
- Theme
- lightweight and on-device AI runtimes
- Vertical
- Horizontal
- Function
- Model & infra
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Commercial product
- Normalized one-liner
- faster llm inference runtime
- Manually corrected
- False
Could you build this?
No BaseRT is a high-performance LLM runtime competing with llama.cpp and MLX, requiring low-level C++/Metal GPU kernel optimization and deep systems engineering for Apple Silicon.
What it would actually take: Developing BaseRT requires custom Metal Shading Language (MSL) compute kernels optimized for Apple's unified memory architecture, AMX (Apple Matrix Coprocessor), and SIMDgroup operations. The engine implements custom matrix multiplication routines, fused attention kernels (e.g., FlashAttention variants for Metal), and custom quantizers (4-bit/8-bit). This level of extreme prefill and decode throughput demands deep expertise in computer architecture, GPU profiling, and assembly-level compiler intrinsics.
Competitors
Other products that read as similar to this one — 19 launches clear the similarity bar, closest 8 shown.
Attention rank: #10 of 20 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 243 days after the earliest competitor.
- Rapid-MLX · hn · 2026-04-18 · 9 upvotes · similarity 0.41
- Ollama v0.19 · ph · 2026-04-01 · 412 upvotes · similarity 0.40
- OS Megakernel that match M5 Max Tok/w at 2x the Throughput on RTX 3090 · hn · 2026-04-08 · 6 upvotes · similarity 0.40
- On the edge of Apple Silicon memory speeds · hn · 2026-01-17 · 5 upvotes · similarity 0.39
- Apple-Watch-Edge-AI · github · 2026-09-16 · 11 upvotes · similarity 0.38
- kvmem-llama.cpp · github · 2026-09-14 · 176 upvotes · similarity 0.35
- bonsai-apple-silicon · github · 2026-09-18 · 24 upvotes · similarity 0.35
- BestLLMfor · ph · 2026-09-09 · 2 upvotes · similarity 0.35
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a model & infra tool for Fintech yet.