Nicheloom visitors: 12,364 all-time · 7,526 in the last 7 days

The opportunity tracker for new startups.

Today: 970 launchs · Niche of the day: niche social and lifestyle utilities · See full report →

Kolibri

Open-weight English-German language model, Apache 2.0

Get picks like this daily. The day's top launches, AI/tech news, and a weekly opportunity spotlight — straight to your inbox.

This is 1 of 48 launches in generative video model API gateways — see how it stacks up on momentum and crowding →

20 other launches read as similar to this one →

Details

External ID
1268591
Source
PH
Company
—
Product
Kolibri
Website domain
producthunt.com
Launched
Oct. 7, 2026
Cohort
—
Upvotes
1
Upvotes percentile
0.33
Tags
Developer Tools, Artificial Intelligence

Description

Kolibri is an English-German Mixture-of-Experts language model from Aleph Alpha, with 78B total and 3B active parameters and a context of up to 1M tokens. Aleph Alpha released it on the Day of German Reunification, with full weights on Hugging Face under the Apache 2.0 license. It is specialized for German, reasoning, math and agentic behavior, aimed at regulated sectors such as public administration, industrials and aerospace, and it is designed to run on-premise. It follows Kolibri Origin, a 3

Enrichment

Niche
generative video model API gateways
Vertical
Horizontal
Function
Model & infra
Audience
Developer
AI stance
AI-native
Project type
Hobby / open-source project
Normalized one-liner
open-weight english-german llm for developers
Manually corrected
False

Could you build this?

No Training a 78B Mixture-of-Experts foundation model with a 1M token context window demands millions of dollars in compute, custom distributed training clusters, and frontier ML research expertise.

What it would actually take: Building Kolibri requires orchestrating thousands of high-end GPUs connected via high-speed InfiniBand using distributed frameworks like Megatron-LM or DeepSpeed. The team must assemble and curate multi-terabyte English and German pre-training corpora, implement custom MoE routing architectures, and tune long-context attention kernels. This requires an elite team of ML research scientists and high-performance computing infrastructure engineers.

Competitors

Other products that read as similar to this one — 20 launches clear the similarity bar, closest 8 shown.

Attention rank: #12 of 21 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).

Launched 258 days after the earliest competitor.

Other launches for this product