Anthropic IPO Valuation Slumps to $80B
AI market rout deepens with valuations revised downward as Anthropic targets $100B run rate.
1
Stories per day · 30 days
8 RSS
Relatedopen-weights8huggingface3meta3multimodal3nvidia3anthropic2inference2inkling2llm2long-context2muse-spark-open2nemotron2agentic1agents1b2001benchmarks1coder1deepseek1
50 stories, newest first · one card per story
Stories tagged #moe
MoE Model Routing Compared: DeepSeek Qwen Mixtral Llama 4 gpt-oss
Analysis details token routing differences across leading Mixture of Experts models.
Thinky Machines releases Inkling 2 Small, 594B MoE beating Muse Spark
Apache 2.0 weights ship next week after safety testing; ranks 31 on Artificial Analysis Index.
SAEs reveal fine-grained specialization in MoE models
Sparse autoencoders trained on Qwen3-30B-A3B show distinct experts for specific tasks like Java coding and geometry.
NVIDIA Shows 3 Ways to Speed Up MoE Inference on DGX Spark, Lifting Ling-3.0-flash Throughput 42.5%
NVIDIA demonstrates three methods to accelerate Mixture-of-Experts inference, increasing Ling-3.0-flash throughput by 42.5% on DGX Spark.
Sonar-MoE Enables Memory-Efficient MoE Training
Sonar-MoE updates expert tensors in the backward pass to cut peak memory by 19-22% for MoE models.
DeepSeek-V4.1-Flash-1017 Coder GGUF model released
A 763B-parameter MoE model pruned to 72.1 GB and 1.05 bits per parameter, optimized for code and agentic workflows.
Scantist/GLM-5.3-SLIM-GGUF 180GB MoE model trends on Hugging Face
179.9GB version of 753B-parameter GLM-5.3 MoE drops to 563B by NAS, scores 92.7% HumanEval and fits single B200.
1
Moonshot AI releases Kimi K3.1 Pro open-weight multimodal model
Kimi K3.1 Pro is an open-weight MoE model with 1M token context and 104B activated parameters.
1
Unofficial Qwen3.8-Flash-Next E384 FP4 model trends on Hugging Face
The NAS build keeps 384 of 512 experts, cuts weights about 16%, and was merged after GRPO-style LoRA.
[AINews] High Taste Testers Ep. 1
AINews covers Meta’s Muse Spark Open and Reflection Beam 1.0 alongside Gemini, Anthropic, and OpenAI updates.
15
StepFun Releases Step 5 Preview Open Weights: 600B MoE, 27B Active, 1M Context
StepFun released Step 5 Preview open weights: a 600B MoE with 27B active parameters and 1M context.
[AINews] Agents disprove Polysemanticity; greatly improve Interpretability & Alignment
Meta released Muse Spark Open, an 816B MoE that leads US open-weight models on Artificial Analysis.
9
Meta Releases Muse Spark Open: 816B Open-Weight MoE With 1M Context
Meta released Muse Spark Open, an 816B open-weight MoE with a 1M-token context.
12
NVIDIA Releases Nemotron 3.5 Super VL, a 120B Open Multimodal Agent Model
NVIDIA released Nemotron 3.5 Super VL, a 120-billion-parameter open multimodal agent model activating 12 billion parameters.
tencent/Hy-Memory-30B-A3B — new model trending #30 on Hugging Face
Tencent released Hy-Memory-30B-A3B, a 3B-active memory writer that leads Mem0, MemOS, and Hy-Memory scores.
Janas: Run LLMs larger than RAM by streaming experts from SSD
Janas streams LLM experts from SSD so models larger than RAM can run locally.
Virtual Head Attention
Virtual Head Attention cuts KV cache versus GQA while improving quality and folding into standard GQA inference.
Meet Cloudflare Clef-omni: A 30B-A3B Multimodal Decision Model for Audio, Video, Image and Text
Cloudflare released Clef-omni, a 30B-A3B Apache-2.0 model that scores text, image, audio, and video.
autotrust/JEV-35B-NVFP4 — new model trending #21 on Hugging Face
AutoTrust released JEV-35B-NVFP4, a 4-bit quantization of its 35B decision model that cuts GPU memory by 65%.
One Model to Loop Them All: Elastic Latent Recurrence for Language Models
Elastic Latent Recurrence trains looped language models to match standard accuracy at one loop and improve with extra loops.
autotrust/GLM5.3-Flash-E224-DGX-Spark — new model trending #29 on Hugging Face
autotrust released an unofficial pruned NVFP4 GLM-5.3-Flash build that fits two NVIDIA DGX Sparks.
Introducing Mistral Large 4: Le chonk
Mistral previewed Mistral Large 4, a 1-trillion-parameter model with 49 billion active parameters.
openbmb/MiniCPM-V-4.7-35B-A3B-gguf — new model trending #30 on Hugging Face
OpenBMB released GGUF weights for MiniCPM-V 4.7 35B-A3B, an on-device multimodal MoE model.