Discussions
What people said about open models on Hacker News and YouTube, each post with the models it is about.
- Kimi K3: Open Frontier Intelligence

Kimi-K3vincent_s · Jul 16, 20262.1K points
- Google releases Gemma 4 open models

gemma-4-31B-it

gemma-4-26B-A4B-itjeffmcjunkin · Apr 2, 20261.8K points
- Qwen 3.8 27B

Qwen3.8-27Berdaltoprak · Aug 14, 20261.4K points
- Kimi-K3 on HuggingFace

Kimi-K3nateb2022 · Jul 27, 20261.4K points
- GLM-5.3: Frontier coding with emergent cyber capabilities
GLM-5.3pella · Aug 14, 20261.2K points - GLM-5.3-Flash
GLM-5.3-FlashPhilpax · Aug 26, 20261.1K points - GLM 5.2 beats Claude in our benchmarks
GLM-5.2jms703 · Jun 28, 20261.1K points - Kimi K3.1 Pro and K3.1 Pro Highspeed

Kimi-K3.1-ProJowsey · Oct 23, 20261.1K points
- Gemma 4 12B: A unified, encoder-free multimodal model

gemma-4-31B-it

gemma-4-26B-A4B-itrvz · Jun 3, 20261.1K points
- DeepSeek V4 Pro 0813

DeepSeek-V4-Proexplosion-s · Aug 12, 20261K points
- DeepSeek v4.1 Flash

DeepSeek-V4.1-FlashLiwink · Sep 10, 20261K points
- DeepSeek-v3.2: Pushing the frontier of open large language models [pdf]

DeepSeek-V3.2pretext · Dec 1, 2025982 points
- DeepSeek-v4.1-Flash-1017

DeepSeek-V4.1-Flashtheykk · Oct 17, 2026959 points
- Show HN: Getting GLM 5.2 running on my slow computer
GLM-5.2vforno · Jul 9, 2026937 points - Show HN: Open-source engine running Gemma 4 26B in 2 GB RAM on any M-series Mac

gemma-4-31B-it

gemma-4-26B-A4B-itgitpusher42 · Jul 29, 2026919 points
- GLM-5.2 is the new leading open weights model on Artificial Analysis
GLM-5.2himata4113 · Jun 17, 2026916 points - Kimi K3 Is Competitive with Fable; Kimi K3 and Fable Is SoTA

Kimi-K3piotrgrabowski · Jul 21, 2026877 points
- Gemma 4 on iPhone

gemma-4-31B-it

gemma-4-26B-A4B-itjanandonly · Apr 5, 2026868 points
- GLM-5.3 is now open-weight
GLM-5.3jeudesprits · Aug 28, 2026806 points - DeepSeek V4 Flash 0731

DeepSeek-V4-Flash-0731

DeepSeek-V4-Flashtosh · Aug 7, 2026796 points
- GLM 5.2 Is Out
GLM-5.2aloknnikhil · Jun 13, 2026772 points - DeepSeek-V4-Flash Update

DeepSeek-V4-Flash-0731

DeepSeek-V4-Flashdnhkng · Jul 31, 2026745 points
- Qwen3-TTS family is now open sourced: Voice design, clone, and generation

Qwen3-0.6B

Qwen3-14BPalmik · Jan 22, 2026744 points
- Qwen3-Coder-Next

Qwen3-Coder-Nextdanielhanchen · Feb 3, 2026735 points
- Qwen3.8-2.4T

Qwen3.8-27B

Qwen3.8-2.4T-A95BPhilpax · Aug 12, 2026713 points
- Kimi K2.6: Advancing open-source coding

Kimi-K2.6meetpateltech · Apr 20, 2026710 points
- Qwenputer: A functioning, deterministic RISC-V CPU inside Qwen3.5

Qwen3.5-2B

Qwen3.5-27Bcsmantle · Oct 29, 2026709 points
- 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 - GLM 5.2 and the coming AI margin collapse
GLM-5.2martinald · Jul 6, 2026694 points - Accelerating Gemma 4: faster inference with multi-token prediction drafters

gemma-4-31B-it

gemma-4-26B-A4B-itamrrs · May 5, 2026687 points
- DeepClaude – Claude Code agent loop with DeepSeek V4 Pro

DeepSeek-V4-Proalattaran · May 3, 2026678 points
- Kimi K3.1 Pro model weights are out

Kimi-K3.1-Proabrarthought · Nov 4, 2026641 points
- The Kimi K3 Moment

Kimi-K3sbochins · Jul 18, 2026636 points
- MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second
MiMo-V2.5-Progainsurier · Jun 8, 2026628 points - GLM-5.1: Towards Long-Horizon Tasks
GLM-5.1zixuanlimit · Apr 7, 2026618 points - GLM-5.2 – How to Run Locally
GLM-5.2TechTechTech · Jun 22, 2026617 points - DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

DeepSeek-V4-Flash-0731

DeepSeek-V4-Flashtheanonymousone · Jul 31, 2026594 points
- GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2
GLM-5.2oshrimpton · Jun 19, 2026585 points - DeepSeek: DeepSeek v4.1 Flash 1017

DeepSeek-V4.1-Flashtheanonymousone · Oct 17, 2026520 points
- GLM 5.2 vs. Opus
GLM-5.2ritzaco · Jun 22, 2026519 points - ZCode – Harness for GLM-5.2
GLM-5.2chvid · Jul 1, 2026511 points - Kimi K3 Architecture Overview and Notes

Kimi-K3ModelForge · Jul 28, 2026507 points
- Kimi Released Kimi K2.5, Open-Source Visual SOTA-Agentic Model

Kimi-K2.5nekofneko · Jan 27, 2026502 points
- Mistral Medium 3.5

Mistral-Medium-3.5-128Bmeetpateltech · Apr 29, 2026500 points
- Kimi K3-256k

Kimi-K3monneyboi · Jul 29, 2026492 points
- How to run Qwen 3.5 locally

Qwen3.5-2B

Qwen3.5-27BCuriositry · Mar 7, 2026490 points
- GLM-5: Targeting complex systems engineering and long-horizon agentic tasks
GLM-5CuriouslyC · Feb 11, 2026484 points - Indexing a year of video locally on a 2021 MacBook with Gemma4-31B (50GB swap)

gemma-4-31B-it

gemma-4-26B-A4B-itasenna · May 21, 2026471 points
- Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hard

Qwen3.8-27Bguregodevo · Oct 31, 2026465 points
- Kev: Tiny Jev-like family of decision models built on top of Qwen3.5

Qwen3.5-2B

Qwen3.5-27Btosh · Sep 21, 2026463 points
- Qwen3.5 122B and 35B models offer Sonnet 4.5 performance on local computers

Qwen3.5-2B

Qwen3.5-27Blostmsu · Feb 28, 2026461 points
- GLM-4.7: Advancing the Coding Capability
GLM-4.7pretext · Dec 22, 2025437 points - Qwen3.5: Towards Native Multimodal Agents

Qwen3.5-2B

Qwen3.5-27Bdanielhanchen · Feb 16, 2026434 points
- Qwen3.5 Fine-Tuning Guide

Qwen3.5-2B

Qwen3.5-27Bbilsbie · Mar 4, 2026416 points
- Kimi K3, and what we can still learn from the pelican benchmark

Kimi-K3droidjj · Jul 17, 2026407 points
- Running Gemma 4 locally with LM Studio's new headless CLI and Claude Code

gemma-4-31B-it

gemma-4-26B-A4B-itvbtechguy · Apr 5, 2026407 points
- Gemma 4 QAT models: Optimizing compression for mobile and laptop efficiency

gemma-4-31B-it

gemma-4-26B-A4B-ittheanonymousone · Jun 5, 2026406 points
- DeepSeek V4 Pro beats GPT-5.5 Pro on precision

DeepSeek-V4-Proyogthos · Jun 8, 2026397 points
- OpenAI says its new model GPT-2 is too dangerous to release (2019)
gpt2surprisetalk · Apr 8, 2026395 points - Show HN: Llama 3.1 70B on a single RTX 3090 via NVMe-to-GPU bypassing the CPU

Llama-3.1-8B-Instruct

Llama-3.1-8Bxaskasdf · Feb 21, 2026395 points