# Kolibri by Aleph Alpha Hacker Approved page: https://hackerapproved.com/p/kolibri/ 21 people recommended Kolibri by Aleph Alpha in Hacker News comments, 9 had mixed feelings and 11 criticized it, November 2025 to October 2026. It was recommended in 2 of 12 months, in 2 threads. It does not meet the Hacker Approved badge rule. 18 of the 21 people who recommend it commented in one thread, “Kolibri: A Sovereign Open-Weight Model”. **Verdict:** Open-weight Apache 2.0 model from Aleph Alpha that people praise for openness, offline-friendly use and strong benchmarks. Critics say it trails Qwen models, is heavy to run locally, and overthinks. Summary written by Claude from 45 comments. It was written before the November 2026 recount, and 42 of those comments still count. ## Why people recommend it - Open weights under Apache 2.0 (3 comments in the sample): "the depth of the tech report shows that they a serious about openness" ([source](https://news.ycombinator.com/item?id=49947937)) "That's the beauty of Apache 2.0! People can just serve it" ([source](https://news.ycombinator.com/item?id=49974722)) "Thank you Aleph Alpha team for making it open." ([source](https://news.ycombinator.com/item?id=49944434)) - Works well in everyday testing (3 comments in the sample): "It's been surprisingly good at the things I threw at it" ([source](https://news.ycombinator.com/item?id=49946290)) "Im surprised by how well this works." ([source](https://news.ycombinator.com/item?id=49944402)) "That is exactly the capability you need for these kinds of use-cases" ([source](https://news.ycombinator.com/item?id=49951667)) - Strong benchmark results (3 comments in the sample): "The benchmarks are impressive given the problem space they are working in." ([source](https://news.ycombinator.com/item?id=49943936)) "Kolibri performs very well in our evaluation" ([source](https://news.ycombinator.com/item?id=49947100)) "it’s actually a good model that also works well on coding and agentic tasks" ([source](https://news.ycombinator.com/item?id=49946275)) - Offline-first for areas without internet (2 comments in the sample): "we do a lot of work to make an offline first education platform" ([source](https://news.ycombinator.com/item?id=47479436)) "You can use Kiwix, OpenStreetMap and Kolibri as an AI-free equivalent." ([source](https://news.ycombinator.com/item?id=47480044)) ## What people warn about - Trails Qwen models on benchmarks (4 comments in the sample): "The absence of any comparison to Qwen3.8 Flash, another MoE model with a small-ish (6B) number of active parameters, is pretty striking." ([source](https://news.ycombinator.com/item?id=49944340)) "Qwen3.8 27B beats Kolibri 79.9 vs 70.8 in German in Kolibri's harness on Kolibri's benchmark." ([source](https://news.ycombinator.com/item?id=49944803)) "Qwen 3.8 27B beats Kolibri on nearly every single benchmark." ([source](https://news.ycombinator.com/item?id=49945944)) - Needs fp16, too big for consumer hardware (2 comments in the sample): "they only have fp16 which won't fit on my Mac" ([source](https://news.ycombinator.com/item?id=49945363)) "Interesting they recommended high end software without considering quant 4 or 8" ([source](https://news.ycombinator.com/item?id=49943873)) - Overthinks and wastes tokens (2 comments in the sample): "spends way too many tokens on overthinking stuff" ([source](https://news.ycombinator.com/item?id=49943869)) "It can't seem to think things through by itself and then make more tool calls to fetch more information." ([source](https://news.ycombinator.com/item?id=49947162)) - Sovereignty claims undermined by backers (2 comments in the sample): "With Schwarz, Cohere, SAP, NVIDIA and the like as backers society can’t have souvereignty in any meaningful way." ([source](https://news.ycombinator.com/item?id=49947899)) "I think it’s a safe assumption that they’re leaning into “sovereign” because performance is bad." ([source](https://news.ycombinator.com/item?id=49944514)) ## Other open-weight language models people recommend - [Qwen 3.8 27B](https://hackerapproved.com/p/qwen-3-8-27b/): 804 recommend, 242 mixed, 301 criticize - [Z.ai GLM](https://hackerapproved.com/p/glm-5-2/): 750 recommend, 181 mixed, 278 criticize - [Kimi](https://hackerapproved.com/p/kimi-k3/): 652 recommend, 182 mixed, 275 criticize - [Gemma 4](https://hackerapproved.com/p/gemma-4/): 524 recommend, 120 mixed, 166 criticize ## The badge A product earns the Hacker Approved badge when at least 10 people recommend it, recommenders outnumber critics at least 2 to 1 (mixed feelings count toward neither side), it was recommended in at least 4 of the 12 months and in at least 5 different threads, and no single thread supplies more than half of the people who recommend it. Health and money products are never badged. ## Cite as Hacker Approved, "Kolibri by Aleph Alpha", https://hackerapproved.com/p/kolibri/ (21 recommend, 9 mixed, 11 criticize; Hacker News comments, November 2025 to October 2026). Counts updated 2 Nov 2026, with data through October 31, 2026.