LLM
MiniMax M3: Open-weight model with a million-token context challenges proprietary leaders
Chinese AI company MiniMax has released its M3 model, an open-weight AI model featuring top-tier coding performance, a...
The Decoder
LLM · Other Companies
LLM
Nvidia's Nemotron 3 Ultra becomes the smartest open US model, but China still leads
Nvidia's new Nemotron 3 Ultra has been ranked as the most capable open-source AI model from the US according to...
The Decoder
LLM
Best Text-to-Speech TTS Models in 2026: A Benchmark-Based Comparison
This article benchmarks and compares leading text-to-speech models in 2026, evaluating their quality, latency, cost,...
MarkTechPost
LLM
NVIDIA Introduces X-Token: Projection-Guided Cross-Tokenizer KD That Outperforms GOLD by +3.82 Average Points on Llama-3.2-1B
NVIDIA introduces X-Token, a projection-guided cross-tokenizer knowledge distillation technique that improves upon the...
MarkTechPost
LLM
StepFun Releases Step 3.7 Flash: A 198B MoE Vision-Language Model for Coding Agents and Search Workflows
StepFun has released Step 3.7 Flash, a 198 billion parameter mixture-of-experts vision-language model designed for...
MarkTechPost
LLM
Liquid AI Releases LFM2.5-8B-A1B: An On-Device MoE Model With 8.3B Total and 1.5B Active Parameters
Liquid AI released LFM2.5-8B-A1B, a mixture-of-experts (MoE) model with 8.3B total parameters but only 1.5B active...
MarkTechPost
LLM
Perplexity AI Open-Sources Unigram Tokenizer That Achieves 5x Lower p50 Latency Than Hugging Face tokenizers Crate
Perplexity AI has open-sourced an optimized Unigram tokenizer that significantly improves performance compared to...
MarkTechPost
LLM
Amazon builds its own AI production platform and greenlights three AI animated series for Prime Video
Amazon MGM Studios and AWS have launched a GenAI Creators' Fund providing filmmakers access to their proprietary AI...
The Decoder
LLM
ElevenLabs Music v2 promises opera-to-metal transitions without losing musical coherence
ElevenLabs has released Music v2, an AI music generation model capable of seamlessly transitioning between different...
The Decoder
LLM
The AI boom drove Nvidia's yearly Taiwan spending from $15 billion to $150 billion
Nvidia's annual spending with Taiwan-based suppliers, primarily TSMC, has surged from $15 billion to $150 billion due...
The Decoder
LLM
Meet EAGLE 3.1: The Speculative Decoding Algorithm That Fixes Attention Drift in LLM Inference
EAGLE 3.1, a new speculative decoding algorithm developed by the EAGLE team, vLLM, and TorchSpec, addresses instability...
MarkTechPost
LLM
7 Ways to Get So Good at AI, People Will Think You Are AI
The article provides practical tips for mastering AI tools and prompting techniques to become proficient with AI...
Wired AI
LLM
Uber president says AI spending is getting ‘harder to justify’
Uber has exhausted its annual AI budget four months into 2026 and is questioning the return on its investments....
The Verge AI
LLM
Together AI Open-Sources OSCAR: An Attention-Aware 2-Bit KV Cache Quantization System for Long-Context LLM Serving
Together AI has open-sourced OSCAR, an INT2 KV cache quantization method that reduces memory usage by 8× and improves...
MarkTechPost
LLM
AI models often give the right answers but point to the wrong sources
Leading AI models like GPT and Gemini often cite incorrect source passages to support their answers, even when the...
The Decoder
LLM
StepFun Releases StepAudio 2.5 Realtime: An End-to-End Voice Model with Roleplay-Specific RLHF and Paralinguistic Comprehension
StepFun released StepAudio 2.5 Realtime, an end-to-end real-time speech language model with customizable persona...
MarkTechPost
LLM
ByteDance study finds that asking LMMs questions beats making it transcribe text for long document training
ByteDance Seed demonstrates that a 7B parameter model can effectively answer questions on long, image-heavy documents...
The Decoder
LLM
Cloudflare CEO Prince says builders and sellers are safe but AI is coming for the measurers
Cloudflare CEO Matthew Prince laid off over 20% of the workforce, attributing the cuts to AI replacing middle...
The Decoder
LLM
Google checks websites for llms.txt in new agentic browsing audit
Google is testing how well websites handle AI agents through a new experimental category called "Agentic Browsing" in...
The Decoder