Explore tools, prompts, skills, news and resources related to 32K.
Prompt works related to 32K
Latest articles, prompts, use cases and editorial content about 32K
Perplexity released pplx-decider-v1.3 and pplx-decider-flash-v1.3, open-weight decision models that return option probabilities instead of generated text for yes/no, multiple-choice, score and locate questions over text...
A new arXiv study finds that storing cached recurrent state in bf16 with round-to-nearest makes hybrid reasoning models more verbose and increases 32K-token cap hits. Across six open hybrid models and 590 competition-mat...
Zyphra released ZYFA, an open-source FP8 and BF16 prefill attention kernel library for AMD Instinct MI355X GPUs. It achieves 1.16-1.76x speedups over AMD's AITER on real model activations, with lower FP8 error. The...
Thinking Machines released Inkling 2 Small, a 594B-total/17B-active multimodal MoE, trained from scratch in 68 hours on 8.4T tokens (NVFP4). It scores 31 on the Artificial Analysis Intelligence Index, beating many same-s...
The roundup covers OpenAI's autonomous security pipeline, which it says validated and generated patches for 948 Linux kernel local privilege-escalation vulnerabilities at a $1.2M API cost. It also reports Moonshot r...
Multiple groups report that frontier LLMs can generate or edit small networks' weights directly, without gradient training or tools. Replications include Alex Krentsel's reproduction of MIT's experiment, H...
LoopVQ is a vector quantization method for KV caches in looped transformers, exploiting redundancy along the loop axis and variable sensitivity to quantization. It achieves 0.5 bits per element, outperforming existing KV...
NVIDIA's technical blog details three optimizations for mixture-of-experts (MoE) inference on a single DGX Spark desktop system, using Ant Group's Ling-3.0-flash model. The optimizations—Humming MXFP4 MoE kerne...
Zyphra released ZHA, a sparse attention method that uses a multi-level tree indexer to select relevant KV cache entries without scanning the full context. The method adds only 81,920 trainable parameters to frozen pretra...
Baseten's Base Labs released joey-1, a 2B parameter decision model that returns labels and calibrated probabilities instead of text. It achieves 20 ms P50 latency at $0.01 per million input tokens, making it suitabl...
Mirendil has released Fern 0.1, its first AI model, a from-scratch coding model with a context window of up to 10M tokens, roughly 23 MB of code. Built on Mirendil-Base, which was pretrained by one researcher with AI age...
Liquid AI introduced LFM3, a family of six open-weight 2B-class models built on one tri-hybrid backbone for text, vision, audio, omni and decision tasks. The 1.7B text model runs in about 1GB at 4-bit with a 256K context...
Hugging Face Bio has released Carbon-X, an Apache 2.0 decoder-only model with 14.8B parameters that processes raw DNA and English in a single token stream and answers in natural language. Built on Qwen3-14B with a hybrid...
Discover tools, products and platforms related to 32K