ByteDance released model with 512K token context
ByteDance company released an open AI model with incredible context of 512,000 tokens. The model name is Seed-OSS-36B. Link in description.
While the world discusses TikTok and the White House, ByteDance quietly rolls out technology that processes information volume equivalent to an entire bookshelf in one session! 3 model versions — with synthetic data, without them, and instructional version — each tailored for its tasks.
The architecture impresses with its elegance. 36 billion parameters distributed across 64 layers. Vocabulary of 155,000 tokens. But the main magic — the thinking budget mechanism! You literally set how much time the model should think before answering. Want instant response — set 0. Need deep analysis — increase the budget.
Test results are awesome! Mathematics — 91.7% on AIME. Programming — 67.4% on LiveCodeBench. Long context work — 94.6% on RULER. All indicators — absolute records among open models!
The key question here is what’s the performance on real tasks, not benchmarks. But so far, ByteDance unexpectedly demonstrates world-class competencies in LLM. This is interesting.
Почитать из последнего
UBTech will send Walker S2 robots to serve on China's border for $37 million
Chinese company UBTech won a contract for $37 million. And will send humanoid robots Walker S2 to serve on China's border with Vietnam. South China Morning Post reports that the robots will interact with tourists and staff, perform logistics operations, inspect cargo and patrol the area. And characteristically — they can independently change their battery.
AI chatbots generate content that exacerbates eating disorders
A joint study by Stanford University and the Center for Democracy and Technology showed a disturbing picture. Chatbots with artificial intelligence pose a serious risk to people with eating disorders. Scientists warn that neural networks hand out harmful advice about diets. They suggest ways to hide the disorder and generate "inspiring weight loss content" that worsens the problem.
OpenAGI released the Lux model that overtakes Google and OpenAI
Startup OpenAGI released the Lux model for computer control and claims this is a breakthrough. According to benchmarks, the model overtakes analogues from Google, OpenAI and Anthropic by a whole generation. Moreover, it works faster. About 1 second per step instead of 3 seconds for competitors. And 10 times cheaper in cost per processing 1 token.
