💬 GPT-5.5 Is Out: Better And Pricier

OpenAI has released its new GPT-5.5 model for paid subscribers. According to the developers, it's the best AI out there for coding, scientific research, work, and agentic tasks.

Benchmarks and independent evaluations back the claims. GPT-5.5 scored 60 points on Artificial Analysis's AI Intelligence Index, edging out its closest rivals, Claude Opus 4.7 and Gemini 3.1 Pro.

💰 Still, GPT-5.5 costs twice as much as the previous version. API pricing sits at $5/$30 per 1M input/output tokens. That's the highest price tag among frontier models, now even steeper than Opus 4.7. On the flip side, the model is very efficient with token usage, which somewhat softens the blow.

💡 Experts at Artificial Analysis also flagged a relatively high hallucination rate in their tests—despite the model having the strongest knowledge base of any available AI. GPT-5.5 makes up answers instead of admitting it doesn't know them about 1.5–2x more often than Opus 4.7 and Gemini 3.1 Pro.

🔍 The frontier labs' release cycle is only halfway through. Elon Musk's xAI is already testing Grok 4.3, and Chinese lab DeepSeek is expected to unveil the long-awaited V4 soon. So "the smartest model from OpenAI" may have some serious new competition right around the corner.

Do you use OpenAI models?

❤️ — Absolutely, they're the best
🔥 — No, switched to Claude/Gemini/Grok/...

@hiaimediaen