Tag

benchmarks

2 posts tagged with "benchmarks"

November 24, 2025

Claude Opus 4.5: Anthropic Reclaims the Coding Crown with 80.9% SWE-bench Score

Anthropic releases Claude Opus 4.5, achieving industry-leading 80.9% on SWE-bench Verified, 66% price reduction, and breakthrough efficiency with the new effort parameter. Here's what developers need to know.

aianthropicclaudecodingbenchmarks

November 20, 2025

Google Gemini 3: Achieving Benchmark Dominance Across AI Leaderboards

Google's Gemini 3 has emerged as a powerhouse, topping 19 of 20 benchmarks and achieving a stunning 2-3x advantage on ARC-AGI 2 over GPT-5.1. Here's what this means for the AI landscape.

aigooglebenchmarksgeminiagi