Trending

Addendum to GPT-5 system card: GPT-5-Codex

GPT-5-Codex

RustGPT: A pure-Rust transformer LLM built from scratch

GPT‑5-Codex and upgrades to Codex

Language Models Pack Billions of Concepts into 12,000 Dimensions

SpikingBrain 7B – More efficient than classic LLMs

Energy-Based Transformers Explained [video]

I unified convolution and attention into a single framework

K2-Think: A Parameter-Efficient Reasoning System

Qwen3-Next 80b MLX (Mac) runs on latest LM Studio

Qwen-next - no gguf yet

[D] How to best fine-tune a T5 model for a Seq2Seq extraction task with a very small dataset?

[D] Paged Attention Performance Analysis

Look at our boi go! WEBGEN-SMALL is a 4B model...

GPT-OSS-120B benchmarks show interesting trade-offs across providers

vLLM is kinda awesome

Qwen3‑Next‑80B‑A3B‑Instruct (FP8) on Windows 11 WSL2 + vLLM + Docker (Blackwell)

Will we see: Phi-5, Granite 4, Gemma 4, Deepseek R2, Llama 5, Mistral Small 4, Flux 2, Whisper 4?

BAIDU

Save item

ERNIE-4.5-21B-A3B-Thinking

Task: text-generation

Qwen

Save item

Qwen3-Next-80B-A3B-Instruct

Task: text-generation

Qwen

Save item

Qwen3-Next-80B-A3B-Thinking

Task: text-generation

LLM360

Save item

K2-Think

Task: text-generation

AI at Meta

Save item

MobileLLM-R1-950M

Task: text-generation

Microsoft

Save item

VibeVoice-1.5B

Task: text-to-speech

inclusionAI

Save item

Ling-mini-2.0

Task: text-generation

Center for Language and Speech Processing @ JHU

Save item

mmBERT-base

Task: fill-mask

Alpha-VLLM

Save item

Lumina-DiMOO

Task: any-to-any