Anthropic Unveils Claude 4.5 with Breakthrough Reasoning Capabilities
TechCrunch
by Sarah Chen
February 11, 2026 · 21:59
5 hours ago
Summary
Anthropic releases its most capable model yet, demonstrating significant advances in multi-step reasoning, code generation, and safety alignment.
Anthropic announced today the release of Claude 4.5, marking a significant leap forward in large language model capabilities. The new model demonstrates marked improvements in multi-step reasoning tasks, mathematical problem solving, and code generation, while maintaining the company's signature focus on safety and alignment.
In benchmarks shared by the company, Claude 4.5 outperformed previous versions across all major evaluation categories, with particularly notable gains in graduate-level reasoning tasks where it achieved a 94.2% accuracy rate.
The release comes amid intensifying competition in the AI industry, with major players racing to develop increasingly capable models while grappling with questions of safety, regulation, and societal impact.
In benchmarks shared by the company, Claude 4.5 outperformed previous versions across all major evaluation categories, with particularly notable gains in graduate-level reasoning tasks where it achieved a 94.2% accuracy rate.
The release comes amid intensifying competition in the AI industry, with major players racing to develop increasingly capable models while grappling with questions of safety, regulation, and societal impact.
Analysis
Relevance Score
Critical
Sentiment
↑ +0.65
Positive
Source Reliability
7.5
/ 10
Medium
Source
TechCrunch
techcrunch.com
Tech
US
Related
Ars Technica · 8 hours ago
Google DeepMind Achieves Protein Folding Prediction at Atomic Scale
The Verge · 10 hours ago
Microsoft and Google Race to Embed AI Agents Directly in Operating Systems
The Information · 11 hours ago
OpenAI Internal Tensions Rise Over Commercialization Speed
Associated Press · 15 hours ago
US Expands Semiconductor Export Controls to Include AI Chip Peripherals