AI Glossary
What is Reasoning model?
Reasoning models spend extra inference-time tokens working through a problem step by step — exploring, backtracking, and verifying — before producing a final answer. This dramatically improves math, science, and code performance at the cost of latency and tokens. They are billed like normal models but consume far more output.
News momentum
Mentions in recent AI news titles and summaries, refreshed from the ingestion stream.
1
-7 vs prior 7d
- Learning When to Think: Adaptive Reasoning for Test-Time Compute Allocation
arxiv-ai · 6d
- FinRCA-Bench: Benchmarking Evidence Retrieval and Reasoning for Financial AI Systems
arxiv-ai · 7d
- Beyond the Trace: Coupling an Interpretable Reasoning-State Readout to Native MoE Routing
arxiv-ai · 8d
- ALPS: Measuring Valid Creativity in Large Language Models with Mathematical Construction
arxiv-ai · 9d