{"rewrite":{"id":"r_4290a5065b14aff3ea7615fd","clusterId":"c_04ae3937d0c7ad9d885581e4","slug":"meta-readies-watermelon-ai-model-to-close-gap-with-gpt-5-5","model":"deepseek-v4-flash","headline":"Meta Readies Watermelon AI Model to Close Gap With GPT-5.5","summary":"Meta Chief AI Scientist Alexandre Wang revealed that the company's Muse Spark model will be updated under the internal codename Watermelon, which has achieved benchmark scores comparable to OpenAI's GPT-5.5. The update focuses on coding and agent capabilities, with internal tests showing a deliberation mode can improve scores by 8%.","whyItMatters":"Meta is positioning Watermelon as a direct competitor to OpenAI's GPT-5.5, closing a measurable performance gap on coding benchmarks while hinting at further improvements to rival Anthropic's Claude Opus.","webCardHtml":"\u003cp\u003eMeta\u0026#39;s next AI model, internally codenamed Watermelon, is on track to match OpenAI\u0026#39;s GPT-5.5 on key coding benchmarks, according to Meta Chief AI Scientist Alexandre Wang. Wang disclosed the update in a public statement, noting that the Muse Spark model\u0026#39;s name will remain unchanged but its internal codename shifts from Avocado to Watermelon. On SWE-Bench Pro, a coding agent evaluation, the previous Avocado version scored 52.5% against GPT-5.5\u0026#39;s 58.6%; GPT-5.5 also led on Terminal-Bench 2.0. Meta says Watermelon has closed that gap. The model includes a deliberation mode that uses AI agents to refine responses, which internal tests show boosts scores by 8%. When asked when a model comparable to Anthropic\u0026#39;s Claude Opus 4.8-which scores 69.2% on SWE-Bench Pro-would arrive, Wang replied, \u0026#39;Soon, very soon.\u0026#39; Some information suggests Watermelon\u0026#39;s output gains stem partly from increased infrastructure usage rather than pure model improvements.\u003c/p\u003e","blueskyPost":"Meta's next AI model, codenamed Watermelon, matches GPT-5.5 on coding benchmarks. Chief AI Scientist Alexandre Wang says a model rivaling Claude Opus is coming 'very soon.'","twitterPost":"Meta's Watermelon AI model matches GPT-5.5 on coding benchmarks. Chief AI Scientist Alexandre Wang says a model rivaling Claude Opus is coming 'very soon.'","threadsPost":null,"newsletterBlurb":"Meta is preparing an update to its Muse Spark AI model under the internal codename Watermelon, which has achieved scores comparable to OpenAI's GPT-5.5 on coding benchmarks. Chief AI Scientist Alexandre Wang also hinted that a model competitive with Anthropic's Claude Opus is coming soon.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20260706-meta-watermelon/\",\"title\":\"Meta's new AI 'Watermelon' will soon arrive with performance rivaling OpenAI's GPT-5.5\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":3791,"outputTokens":611,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1783347401,"createdAt":"2026-07-06T14:09:25.000Z","publishedAt":"2026-07-06T14:13:32.000Z","updatedAt":"2026-07-06T14:09:25.000Z"},"cluster":{"id":"c_04ae3937d0c7ad9d885581e4","canonicalTitle":"Metaの新AI「Watermelon」はOpenAIのGPT-5.5に匹敵する性能で近日登場","representativeArticleId":"a_d4368c1910aacf604712421e","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[],\"studios\":[],\"people\":[\"Alexandre Wang\"],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-07-06T13:25:00.000Z","lastSeenAt":"2026-07-06T13:25:00.000Z","updatedAt":"2026-07-06T14:13:33.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20260706-meta-watermelon/","title":"Metaの新AI「Watermelon」はOpenAIのGPT-5.5に匹敵する性能で近日登場"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":[],"studios":[],"people":["Alexandre Wang"],"type":"news","domain":"other","is_roundup":false},"keyFacts":null}