Opus 5 is currently #1 on Artificial Analysis Intelligence Leaderboard
Points and comments are a snapshot, not live.
Anthropic's Opus 5 tops the Artificial Analysis Intelligence leaderboard on benchmark scores.
The leaderboard ranks AI models on intelligence, cost, speed, and other metrics using a v4.1 index across nine evaluations, including agentic benchmarks and hallucination tests. Opus 5 currently holds the #1 overall intelligence score. A cost-versus-intelligence chart shows GPT-5.6 Sol Max as cheaper and nearly equivalent on index score. Models can be compared by reasoning effort levels and openness.
What commenters are saying
Commenters debate whether Opus 5's #1 ranking is meaningful, arguing that cost-performance is more important. Several point out that GPT-5.6 Sol and Luna offer comparable intelligence at far lower API costs. Some report Opus 5 failing basic tasks like debugging or getting confused by permissions, while others praise its benchmarks but note censorship and safety restrictions as reliability issues. A few predict DeepSeek v5 will disrupt pricing.