Multimodal reasoning in 2026: see, read, and prove
Frontier systems combined diagrams, PDF papers, and code in one tool-verified reasoning thread.
4 articles
Razonamiento
Index of editorial tags and filtered coverage.
o3 and DeepSeek-R1: the test-time reasoning race
Models optimized to think longer at inference defined 2025, with leaps in math contests and coding agents.
DeepSeek: Chinese open models that shook cost assumptions
DeepSeek released competitive reasoning models with reportedly low training costs, altering market expectations.
OpenAI o1 (2024): models trained to reason longer
The o1 series introduced inference with internal “chain of thought” optimized by RL, excelling at math and code.