In a head-to-head coding test, OpenAI's GPT-5.6 Sol reportedly generated a full, working Minecraft clone from a single prompt, matching a benchmark that rival Anthropic's Claude Fable 5 established at launch. The comparison, using the same task and the same prompt with no retries, has become the latest flashpoint in an intensifying rivalry over how well frontier models can autonomously build large software projects.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.