Arena updated the Image-to-WebDev leaderboard: GPT-6 Astra (Max) leads at 1733 points, 129 above GPT-5.6 Sol (xHigh); Claude Fable 5.1 (Max) is second at 1710. The board compares newly evaluated model releases on image-to-webdev delivery.
Arena.ai says Anthropic Claude Fable 5.1 (Max) is #1 on Agent Arena across 6.7k+ real-world long-horizon agentic sessions with +15.8% net improvement, redrawing the price-performance frontier at a $4.14 median cost/task. Implicit praise-vs-complaint leads by +42.5%, confirmed success +22.4%, bash recovery +13.1%, and no tool hallucinations; GPT-6 Astra traces are still coming in.
CoArena (@coastyai, YC S26) released CUA KnowledgeBench (Real Company Work): a 12-hour horizon with 10 private real-world tasks ร 5 seeds across 100+ apps, scored pass/fail on the work product. After spending over $50k on frontier models, Claude Fable 5.1 solved 12 of 50 twelve-hour runs (24%) while GPT-6 Astra barely cracked the top 3. Next they plan $200k+ compute for 1-, 3-, and 7-day horizons.
Artificial Analysis (@ArtificialAnlys) shipped Intelligence Index v4.3: Terminal-Bench upgraded 2.1โ4.0 (now mini-SWE-agent harness) and ฯยณ-Banking replaced by AutomationBench-AA (657 private Zapier business workflows). Claude Fable 5.1 and GPT-6 Astra tie at 53; OpenAI holds most of the intelligence-vs-cost frontier across Astra reasoning efforts.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.