Arena: Gemini 4 Argon Takes the Lead in the Rankings

The collaborative Arena ranking relies on anonymized duels and an Elo score, where winning against a stronger opponent yields more points. In October 2026, the hierarchy shifts in text-to-text with Gemini 4 Argon ahead of Muse Wonder. By categories, Anthropic excels in document analysis and dominates the podiums, while OpenAI and Qwen remain present in web search and vision.

Arena ranks through anonymized duels and an evolving Elo score

Arena pits two anonymized AI models against each other on the same query, with the user selecting the most relevant response. Each model receives an Elo score, adjusted after each duel: defeating a higher-ranked opponent earns points, while losing to a lower-ranked one results in a loss of points. Even at the top, a model can continue to improve its score, but the rate of progress slows down. This benchmark continuously tracks tasks such as mathematics, coding, or creative writing, and is closely monitored by industry players and observers.

Categories: OpenAI leads in web, Anthropic excels in documents

In the Web Search category, GPT-5.6 Sol "xHigh" holds the top position, followed by Qwen 3.8 Max and Claude Opus 4.6 Search. Among the top ten, nine spots are shared between Anthropic (five) and OpenAI (four). For Image Analysis, Claude Fable 5 comes in first, ahead of Qwen 3.8 and Claude Opus 4.6, with Anthropic placing five models in the top 10. In Document Analysis, Claude Opus 5 ranks first, and Anthropic claims the top seven positions; OpenAI also appears in the top 10 with GPT-5.5 standard, GPT-5.5 "High," and GPT-5.6 Sol.

Text-to-text: Google takes the lead, Anthropic remains omnipresent

In October 2026, Gemini 4 Argon rises to the top of the text-to-text ranking, ending several months of Anthropic's dominance. This model, recently offered by Google to its paying subscribers, surpasses Muse Wonder. Anthropic maintains a strong presence with six models in the top half of the table. The top 20 text-to-text includes, in order, Gemini 4 Argon "High," Muse Wonder "Max," Claude Opus 5.5 "High," Claude Fable 5 "High," Claude Opus 4.6 "High," Claude Fable 5.1 "High," Claude Opus 4.7 "High," Claude Opus 4.6, Gemini 3.8 Flash "High," Muse Spark 1.3 "Max," Claude Opus 4.7, Muse Spark 1.2 "xHigh," Muse Spark 1.1, Claude Opus 5 "Max," Claude Opus 5 "High," Muse Spark, Gemini 3.7 Flash "High," Kimi K3 "Max," Gemini 3.1 Pro Preview, and Gemini 3 Pro. With the exception of Moonshot, which places Kimi K3 in 18th position, the top twenty spots are occupied by Meta, Google, and Anthropic.