Arena tracks frontier models' Ancient Rome progress; Claude Sonnet 5.5 and GPT-6.1 scores coming
arena · x · 2026-09-30
Arena showcased how frontier models have progressed on bringing Ancient Rome to life from Q4 2025 to now, featuring Claude Opus 4.5–5.5, Claude Fable 5/5.1, GPT-5.2 through GPT-6 Astra/Sol, Gemini 3.1, Kimi K3, and Qwen 3.8 Max. Scores for Claude Sonnet 5.5 and GPT-6.1 are coming soon, powered by real-world tasks from its global user community.
More from Multimodal
- Cardboard claims it has "solved video editing once and for all" — free on Mac — pritisinghhhh · 2026-09-30
- Two weeks learning K2T and MMH3 unlocked consistent characters across models and LoRAs — Valuable-Mango-6710 · 2026-09-30
- Free Midjourney Access via Meta AI, With ChatGPT-Written Prompts — Feeling_Shake1538 · 2026-09-30
- 30-second sneaker motion film made entirely in Hailuo AI from a single image — LudovicCreator · 2026-09-30
- Midjourney v8.2 demo: changing one word dramatically alters the output — tisch_eins · 2026-09-30
- Community LoRA fixes high-speed flip and spin motion breaks in MiniMax-H3 — fruesome · 2026-09-30