Discussions
What people said about open models on Hacker News and YouTube, each post with the models it is about.
- Calibri runs 744B parameter AI on laptop without a GPU using GLM 5.2 #CalibriAI #GLM5 #AIDisruption
GLM-5
GLM-5.2Jordan Blake · Sep 12, 202659.4K likes - GLM-5.3-Flash
GLM-5.3-FlashPhilpax · Aug 26, 20261.1K points - За считанные минуты локальная ИИ модель моделирует корпус для ESP32C3 #ai #qwen3

Qwen3-0.6B

Qwen3-14BУ Павла! · Oct 27, 202612.3K likes
- Kimi K3.1 Pro and K3.1 Pro Highspeed

Kimi-K3.1-ProJowsey · Oct 23, 20261.1K points
- I ran Qwen3.8 27B on single 8GB Card! (Better than expected)

Qwen3.8-27BRed Stapler · Aug 24, 20267.9K likes
- What It Actually Costs to Run DeepSeek V4.1 Flash Locally?

DeepSeek-V4.1-FlashKai · Sep 13, 20266.6K likes
- DeepSeek v4.1 Flash

DeepSeek-V4.1-FlashLiwink · Sep 10, 20261K points
- Локальная Qwen3.8:27b вместе с Hermes удивляет! #llm #ollama #hermes

Qwen3.8-27BУ Павла! · Sep 23, 20266.5K likes
- DeepSeek-v4.1-Flash-1017

DeepSeek-V4.1-Flashtheykk · Oct 17, 2026959 points
- GEMMA4 DE GOOGLE - IA EN LOCAL GRATIS EN TU MOVIL#ia #ai #gemini #bigdata #data #bilbao #gemma

gemma-4-31B-it

gemma-4-26B-A4B-itmaxmaxdata · Aug 27, 20266.5K likes
- Qwen3.8 27B is something else..

Qwen3.8-27BCaleb Writes Code · Aug 26, 20264.6K likes
- GLM-5.3 is now open-weight
GLM-5.3jeudesprits · Aug 28, 2026806 points - DeepSeek V4-Flash локально | 288GB VRAM на четырёх GPU

DeepSeek-V4-Flash-0731

DeepSeek-V4-FlashServerFlow - серверы и серверные комплектующие · Aug 20, 20264.4K likes
- DeepSeek-V4.1 Flash: The Most Insane Optimization So Far

DeepSeek-V4.1-Flashbycloud · Oct 10, 20264.4K likes
- Qwenputer: A functioning, deterministic RISC-V CPU inside Qwen3.5

Qwen3.5-2B

Qwen3.5-27Bcsmantle · Oct 29, 2026709 points
- Qwen 3.8 27B локально! Отменяю подписку на Claude?

Qwen3.8-27BСтудия Игор · Aug 27, 20263.9K likes
- I spent $266 and four AI models to own my tablet. GLM-5.3 finished it in a day
GLM-5.3dr_pardee · Aug 23, 2026706 points - 本地模型开发FPS游戏,0元享上亿Token,Qwen3.8‑27B完整实操,VibeCoding教程

Qwen3.8-27BZZ Dream · Sep 2, 20263.7K likes
- GLM 5.3: Powerful AI Is Becoming Almost Free
GLM-5.3Two Minute Papers · Sep 1, 20263.4K likes - Kimi K3.1 Pro model weights are out

Kimi-K3.1-Proabrarthought · Nov 4, 2026641 points
- Run Qwen3.8-27B on ANY GPU (8GB to 32GB): Here's How

Qwen3.8-27BRepoChad · Aug 31, 20263.3K likes
- Best Ways to Get 32GB VRAM for Qwen3.8-27B (Full Context)

Qwen3.8-27BRepoChad · Sep 15, 20263.1K likes
- DeepSeek: DeepSeek v4.1 Flash 1017

DeepSeek-V4.1-Flashtheanonymousone · Oct 17, 2026520 points
- Build & Train a GLM-5.3-Flash Model From Scratch with Python
GLM-5.3-FlashfreeCodeCamp.org · Oct 6, 20262.7K likes - Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hard

Qwen3.8-27Bguregodevo · Oct 31, 2026465 points
- Cancel your subscriptions, Ox-Alpha is here! (GLM 5.3 Flash)
GLM-5.3-FlashMatthew Berman · Aug 29, 20262.5K likes - The New DeepSeek V4.1 Flash Is INSANE – The BEST Open Model Yet?

DeepSeek-V4.1-FlashBijan Bowen · Oct 18, 20262.3K likes
- Kev: Tiny Jev-like family of decision models built on top of Qwen3.5

Qwen3.5-2B

Qwen3.5-27Btosh · Sep 21, 2026463 points
- Qwen3.8 27B: Same Model, Three Harnesses, One Clear Winner

Qwen3.8-27BJames Layne · Aug 28, 20262.2K likes
- 【米中AI覇権競争】中国AI「Kimi K3」実力は本物か、幻か? 米中がAIで世界を二分する/中国AIの優位・劣位/AIの急所を日本が握る《峯村健司×柯隆×興梠一郎》

Kimi-K3楽待 RAKUMACHI · Aug 26, 20262K likes
- Inkling 2 Small
Inkling-2-SmallPhilpax · Nov 6, 2026298 points - 中國Kimi K3免費開源震撼全球AI市場!烏凌翔揭密中國利用MoE專家系統打破縮放定律、獨佔AI蛋糕理論應用層? ft.科技力智庫 烏凌翔【#財經相對論 EP107】CC字幕 @cteevideo

Kimi-K3工商時報 · Aug 27, 20261.8K likes
- LFM2.5-24B-A2B: Built for fast on-device agents
LFM2.5-24B-A2Bgmays · Oct 14, 2026288 points - GPT 6 Astra vs GLM 5.3 vs Gemini 3.8 Criando SHADERS no Minecraft… ficou INSANO 😱
GLM-5.3Caverna da IA · Sep 29, 20261.6K likes - Qwen3.8 27B + DeepSeek Harness BEATS Opus 4.6 in Claude Code

Qwen3.8-27BJames Layne · Aug 21, 20261.6K likes
- Benchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapses

Qwen3.8-27Bstared · Sep 8, 2026286 points
- Qwen3.8 27B大火!实现Token自由?Mac mini实测11个本地模型,表现如何...

Qwen3.8-27BTuTu生活志 · Aug 22, 20261.6K likes
- I Tested DeepSeek V4.1 Flash vs GPT-6 Astra (RAW RESULTS)

DeepSeek-V4.1-FlashBart Slodyczka · Sep 16, 20261.5K likes
- Kimi K3 (2.8T) at 1 token/s on a MacBook Pro, streamed from four SSDs

Kimi-K3Argonautlabs · Sep 8, 2026279 points
- I Tested Deepseek V4.1 Flash vs Qwen 3.8 Flash, GLM 5.3 Flash, Gemini 3.8 Flash
GLM-5.3-Flash
DeepSeek-V4.1-FlashRemakeBench · Sep 18, 20261.4K likes
- Show HN: DeepSeek V3 in ~200 lines of pure Jax

DeepSeek-V3

DeepSeek-V3-0324VaibhaavTiwari · Nov 14, 2026276 points
- Жёсткие тесты: GPT-6 Astra vs Fable 5.1, Opus 5, GPT 5.6 Sol, Kimi K3, GLM 5.3
GLM-5.3
Kimi-K3Олег Стефанов · Sep 8, 20261.3K likes
- GLM 5.3 Flash Is HERE – Is THIS Better Than the FULL GLM 5.3?
GLM-5.3-Flash
GLM-5.3Bijan Bowen · Aug 27, 20261.1K likes - Unsloth Qwen3.8-27B Quantizations – Dynamic v3.1 Update

Qwen3.8-27BRetro_Dev · Nov 14, 2026261 points
- 【ミュトス級のサイバー能力なのに悪用対策が緩い】AIを“ちょっと騙す”だけで攻撃に使える/中国AI「GLM-5.3」の脅威/相次ぐ情報漏洩…GPT・Claude最新モデルの関係は【1on1 Tech】
GLM-5.3TBS CROSS DIG with Bloomberg · Oct 8, 20261.1K likes - 🤯 ना ChatGPT, ना Gemini! Google का ये Secret AI बिना Internet चलेगा 😱 | Gemma 4 FREE!

gemma-4-31B-it

gemma-4-26B-A4B-itmybigguide · Oct 24, 20261K likes
- GLM-5.3 and the spread of advanced cyber capabilities
GLM-5.3Philpax · Sep 29, 2026256 points - Building a gpt2 124M Inference Engine From Scratch | Super30 livestreams
gpt2Harkirat Singh · Sep 15, 2026950 likes - Qwen3-Coder writes less secure code when told it's for a US persona

Qwen3-Coder-30B-A3B-Instruct

Qwen3-Coder-480B-A35B-Instructmattifrank · Oct 21, 2026243 points
- Project Maya Deepseek V4.1 Flash 1017 + GLM 5.3 Local Ai Review
GLM-5.3
DeepSeek-V4.1-FlashDigital Spaceport · Oct 20, 2026830 likes
- This Tiny Coding Model Shouldn’t Be This Good (MiniCPM5)
MiniCPM5-2B
MiniCPM5-1BBetter Stack · Sep 15, 2026766 likes - GLM-5.3 (open-weight) beat Anthropic/OpenAI models – for 1/5 the cost
GLM-5.3ed-is-ai · Aug 23, 2026240 points - MiniCPM5-2B: The Best Sub-Agent Model Yet?
MiniCPM5-2B
MiniCPM5-1BSam Witteveen · Sep 10, 2026702 likes - Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash

DeepSeek-V4-Flash-0731

DeepSeek-V4-FlashHenryNdubuaku · Sep 18, 2026236 points
- Stop paying for CLAUDE or CODEX! Try Nvidia Nemotron 3 Ultra

NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16Data Logic With Shivam · Oct 13, 2026633 likes
- NVIDIA Nemotron 3 Ultra: бесплатная ИИ-модель для кодинга вместо Claude Code и Codex

NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16Ринат Сулейманов | AITron · Oct 27, 2026426 likes
- One month coding with GLM 5.3 Flash
GLM-5.3-FlashThibWeb · Oct 2, 2026234 points - 4 Kata Kadar Hızlı: NVIDIA Nemotron 3.5 Lightning

NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16Ceyhun Tokgezer · Nov 9, 2026351 likes
- NVIDIA lanza nemotron 3 ultra: Open source, gratis y más rápido para programar en tu editor

NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16Hermanas Carvajalino · Sep 22, 2026349 likes
- Five Qwen3.8-27B models match Fable 5 Using GVS5H Harness

Qwen3.8-27Brzk · Oct 25, 2026196 points