Mistral AI Launches Large 3 to Rival GPT-5
French AI startup Mistral AI unveils its most powerful model yet, targeting enterprise workloads with performance it cla…
📂 LLM News
Latest articles in LLM News
OpenAI Upgrades ChatGPT to GPT-5.5 Instant Model
OpenAI replaces ChatGPT's default model with GPT-5.5 Instant, cutting hallucinations by up to 52.5% and delivering more …
VinAI Launches Top Southeast Asian Multilingual LLM
Vietnam's VinAI Research releases a state-of-the-art multilingual language model optimized for Southeast Asian languages…
A 1930s AI With Zero Tech Knowledge Now Writes Code
Researchers fine-tuned an AI trained only on pre-1931 data into a software engineer using just 250 examples, proving LLM…
VinAI Launches Open-Source Multilingual Model for SEA
Vietnam's VinAI releases an open-source foundation model designed for Southeast Asian languages, challenging Western LLM…
NTT Research Launches Tsuzumi 2.0 LLM for Enterprise
NTT Research unveils Tsuzumi 2.0, a Japanese-English bilingual LLM targeting enterprise customers with lightweight archi…
SmolLM 3: Hugging Face's Compact Model Rivals GPT-4o Mini
Hugging Face launches SmolLM 3, a small but powerful open-weight language model that challenges GPT-4o Mini across key b…
Claude 4 Opus Beats GPT-5 in Coding Benchmarks
Anthropic's Claude 4 Opus scores 92.4% on SWE-bench, outperforming OpenAI's GPT-5 by 7 points in software engineering ta…
NTT Unveils Tsuzumi 2.0 LLM for Edge Devices
Japanese telecom giant NTT launches Tsuzumi 2.0, a lightweight large language model designed to run efficiently on edge …
Rakuten Launches 70B Open-Source Japanese LLM
Rakuten releases its largest open-source Japanese language model with 70 billion parameters, challenging Western AI domi…
DeepSeek V4's Biggest Regret: Where Is Engram?
DeepSeek V4 shipped with mHC, CSA, Muon, and FP4 — but the community's most anticipated feature, Engram, is nowhere to b…
Naver HyperCLOVA X Beats GPT-4 on Korean Tasks
South Korea's Naver claims its HyperCLOVA X model outperforms OpenAI's GPT-4 on Korean language understanding benchmarks…