Search results for leaderboard
AI & Enterprise
Xenon releases Hunmin VLM 397B as open source
Xenon said on Thursday it has released as open source Hunmin VLM 397B, an AI model that recognises and directly manipulates computer screens to perform tasks. Built on Qwen3.5-397B, the model scored 75.6 on ScreenSpot-Pro to rank second on Hugging Face\'s official leaderboard, and 70.5 on OSWorld in a Linux environment, 22.3 points above the base model. The company trained it using eight Nvidia B200 GPUs.
AI & Enterprise
LG AI Research says it will bet on expert AI to solve industry problems
LG AI Research set out an AGI strategy it said differs from global big tech, aiming to keep improving general models while building expert AI that can solve difficult problems in manufacturing, science and finance. It unveiled sector applications and development plans at an event in Seoul. The group is developing EXAONE model series, domain-specific models and K-EXAONE. It also said AI development should continue, while releases and deployment require greater caution.
Crypto
Quantum attack computation cost estimate on bitcoin, ether drops by more than half in two months
Estimates for the core computation cost of a potential quantum-computer attack targeting bitcoin and ethereum have fallen to less than half of previous levels. The work focused on optimising parts of Shor\'s algorithm used to attack secp256k1, a cryptographic scheme shared by both networks. Researchers stressed this does not mean either system can be cracked with current quantum computers, but the faster-than-expected decline in costs is sharpening calls to bring forward security transitions.
-
AI & Enterprise
Anthropic\'s Claude Fable 5.1 rated No.1, beating GPT and Grok
-
AI & Enterprise
Linkalpha hires former Coupang Pay CFO Na Hye-won as COO
-
AI & Enterprise
Mystery AI model Ox Alpha confirmed as China\'s Z.ai product
-
Telecommunications & Media
SKT steps up scale-up of proprietary AI to multimodal and agents
-
AI & Enterprise
SKT\'s A.X K2 proves real-world competitiveness beyond technical capabilities
-
AI & Enterprise
SKT, LG AI Research and Upstage advance to third evaluation of sovereign AI
-
AI & Enterprise
Upstage unveils Solar Pro 4, available on OpenRouter and others
-
AI & Enterprise
DeepSeek releases official V4 Pro model; 46 times cheaper than Claude with 5 percent performance gap
-
AI & Enterprise
Local PC LLMs gain traction as tech firms step up competition
-
AI & Enterprise
SKT\'s A.X K2 posts top-tier results in AI math reasoning evaluations
-
Crypto
Foggd links crypto market maker leaderboard with DefiLlama
-
AI & Enterprise
LG AI Research industrial-focused AI strategy gains traction, stands out on global benchmarks
-
AI & Enterprise
DeepSeek launches coding AI model 99 percent cheaper than Anthropic\'s Claude Opus 4.8
-
AI & Enterprise
Alibaba teases Qwen3.8, says it ranks second only to Anthropic\'s Claude Fable5
-
Industry
Silicon Valley takes cautious look at China’s Kimi K3 despite benchmark lead