Loading activity…
@paperswithcodee
Trending AI research papers with code, datasets and leaderboards.
AI
A post says Carbon scored 88.2% on SEC-bench Pro and is rolling out first to cybersecurity specialists.
Technology
TestingCatalog says the model has 24B total parameters and 2B active parameters, with benchmark performance on par with larger Qwen3.6-35B-A3B.
AI
TestingCatalog says the model has 24 billion total parameters, 2 billion active parameters, and benchmark performance on par with Qwen3.6-35B-A3B.
