OpenAI released GPT-5.6 on June 26 as a limited-preview family of frontier models comprising three tiers named Sol, Terra and Luna, with pricing of $5 per million input tokens and $30 per million output tokens, and the rollout is initially restricted to vetted partners via API pending security and release‑process review. Binance’s 5‑minute LUNA2/USDT futures chart showed the price climbing from about $0.0486 to $0.0513, and open interest rose from roughly 36.5 million to 52.3 million contracts, a 43 % increase, while funding turned positive at 0.01 %. LUNA2, the governance token for Terra 2.0, had a market capitalization near $36 million and 24‑hour trading volume of about $8.5 million, with no spot market activity on Coinbase. Traders reported buying the token because its name matched one of OpenAI’s new model tiers, expecting rapid attention from bots and social media users; a crypto researcher described the behavior as semantic arbitrage, where traders exploit the speed at which a recognizable word spreads through the crypto attention economy. Similar name‑collision trades occurred earlier with TRUMP tokens after a gala invitation announcement, PENGUIN tokens after a viral White House image, and GORK after a tweet by Elon Musk, each driven by keyword adjacency rather than token fundamentals. The increase in open interest outpaced price movement, indicating leveraged positions rather than spot accumulation, and exchanges may adjust margin requirements for tokens that experience sudden open interest spikes, while funding costs can rise quickly in crowded semantic trades, penalizing late entrants. The episode demonstrates that cultural keywords linked to low‑liquidity tokens can produce temporary price dynamics, and traders may continue to monitor AI model names and other cultural references for comparable opportunities.