Arena's Agent Arena ranks Anthropic's Claude Fable 5.2 (Max) #1, ahead of Claude Opus 5.5 (Max) in 9 of 10 categories
Arena's Agent Arena leaderboard places Anthropic's Claude Fable 5.2 (Max) at #1, reporting a +30.5% net improvement across 7.8k real-world sessions as of November 13, 2026. Per Arena, Fable 5.2 leads Claude Opus 5.5 (Max) in 9 of the 10 categories where both are ranked, with a statistical tie in the tenth, and Arena attributes the largest lead to visual design. Arena reports overall net improvement scores of 30.48% for Fable 5.2 versus 19.51% for Opus 5.5, and says the Business & Finance gap is too small to call.
Arena's Agent Arena leaderboard places Anthropic's Claude Fable 5.2 (Max) at #1, reporting a +30.5% net improvement across 7.8k real-world sessions as of November 13, 2026. Per Arena, Fable 5.2 leads Claude Opus 5.5 (Max) in 9 of the 10 categories where both are ranked, with a statistical tie in the tenth, and Arena attributes the largest lead to visual design. Arena reports overall net improvement scores of 30.48% for Fable 5.2 versus 19.51% for Opus 5.5, and says the Business & Finance gap is too small to call.
Key facts
- Arena's Agent Arena leaderboard ranks Anthropic's Claude Fable 5.2 (Max) #1, with a +30.5% net improvement across 7.8k real-world sessions as of November 13, 2026.
- Per Arena, Claude Fable 5.2 (Max) leads Claude Opus 5.5 (Max) in 9 of the 10 Agent Arena categories where both are ranked, with a statistical tie in the 10th.
- Arena reports overall net improvement scores of 30.48% for Fable 5.2 versus 19.51% for Opus 5.5 (both Max).
- Arena reports category net improvement scores of 45.34% vs. 23.29% in visual design, 31.51% vs. 17.88% in documents, 25.67% vs. 13.24% in knowledge work, and 31.45% vs. 22.24% in code.
- Arena says the gap between Fable 5.2 and Opus 5.5 in Business & Finance is too small to call.
- Arena attributes Fable 5.2's largest lead over Opus 5.5 to visual design.
- Arena says the scores only show that Fable is ahead, not where the gap comes from, and that it examined what makes Fable 5.2 different from Opus 5.5 across real-world sessions.
- Arena reports Claude Fable 5.2 (Max) at #1 in Agent Arena with a +30.5% net improvement across 7.8k real-world sessions as of November 13, 2026.
Sources · 1 source
- AA
Arena.ai@arenaPost on X ·
This Week in the Arena: Fable 5.2 takes #1, and Gemini 4 Hydrogen brings the speed. - Claude Fable 5.2 (Max) by @AnthropicAI takes #1 in Agent Arena with +30.5% net improvement across 7.8k real-world sessions as of today. - Gemini 4 Hydrogen (High) by @GoogleDeepMind launched as the fastest model in the Agent Arena top 10, with a typical step of 7.3 seconds. - Muse Ember (Max) by @AIatMeta, Meta’s smallest model, debuted in Code Arena: WebDev. It’s #37 with 1609 points today, at a $0.03 median cost per task. - Inkling 2 Small by @thinkymachines is now open weights and the #1 US open model in Code Arena: WebDev at 1647 points. - MiniMax H3-Image by @MiniMax_AI is the #1 open model in Multi-Image Edit at 1365 (#11 overall), on par with Nano Banana 2 and Nano Banana Pro. - ERNIE 5.5 Flash Preview by @Baidu_Inc is trending toward #2 open model in Code Arena: WebDev at 1668 points (AutoEval), with a 1M-token context window. Open weights under Apache 2.0 are coming soon. https://t.co/wXf4WOlqMA
Open source - AA
Arena.ai@arenaPost on X ·
Claude Fable 5.2 (Max) by @AnthropicAI ranks #1 on Agent Arena, ahead of Claude Opus 5.5 (Max) in 9 of the 10 categories where both are ranked. In the 10th category, there is a statistical tie. Where Fable really shines over Opus is its visual taste. Net improvement score for Fable 5.2 vs. Opus 5.5 (both Max) across key categories: - Overall: 30.48% vs. 19.51% - Visual design: 45.34% vs. 23.29% - Documents: 31.51% vs. 17.88% - Knowledge work: 25.67% vs. 13.24% - Code: 31.45% vs. 22.24% In Business & Finance, the gap is too small to call. But the scores only show that Fable is ahead, not where the gap comes from. So we looked at what makes Fable 5.2 different from Opus 5.5 across real-world sessions.
Open source