{ "generatedAt": "2026-11-13T00:21:39.098Z", "note": "История на изданията на Мерило: един запис на модел за всяка дата в bench/results/. Резултатите са в скала 0–100: „overall“ и „categories“ са автоматичните проверки по факти (C4 там са трудните проверки за писма), „letters“ е независимата оценка на писмата от двамата оценители. „costUsd“ е платената сума за отговорите на модела в изданието, когато е записана (cost.json); за изданието от 15 октомври 2026 г. тя не е известна поотделно. Данните на Основа са CC BY 4.0; сравнението е разработено с помощта на Claude (Anthropic), а модели на Anthropic също участват.", "runs": [ { "date": "2026-10-15", "models": 19, "answers": 3610, "answersUsd": 67.34, "judgingUsd": 16.9, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] } ], "rows": [ { "model": "alibaba/qwen3.8-max-prime", "vendor": "Alibaba (Qwen)", "date": "2026-10-15", "overall": 61.7, "categories": { "C1": 60, "C2": 72.1, "C3": 36, "C4": 90, "C5": 86.7, "C6": 93.3 }, "letters": 74.5, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "anthropic/claude-fable-5.1", "vendor": "Anthropic", "date": "2026-10-15", "overall": 80.6, "categories": { "C1": 96, "C2": 95, "C3": 66, "C4": 85, "C5": 93.3, "C6": 60 }, "letters": 68.1, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "anthropic/claude-haiku-5.5", "vendor": "Anthropic", "date": "2026-10-15", "overall": 58.2, "categories": { "C1": 56, "C2": 67.5, "C3": 36, "C4": 70, "C5": 66.7, "C6": 86.7 }, "letters": 65.2, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "anthropic/claude-opus-5.5", "vendor": "Anthropic", "date": "2026-10-15", "overall": 74.9, "categories": { "C1": 76, "C2": 83.3, "C3": 56, "C4": 90, "C5": 86.7, "C6": 93.3 }, "letters": 68.8, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "anthropic/claude-sonnet-5.5", "vendor": "Anthropic", "date": "2026-10-15", "overall": 68.4, "categories": { "C1": 64, "C2": 78.3, "C3": 50, "C4": 88.3, "C5": 84.4, "C6": 93.3 }, "letters": 66.5, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "bggpt-gemma-3-27b-fp8", "vendor": "INSAIT", "date": "2026-10-15", "overall": 53.6, "categories": { "C1": 84, "C2": 60.4, "C3": 38, "C4": 56.7, "C5": 84.4, "C6": 20 }, "letters": 67.1, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "deepseek/deepseek-v4-pro-0813", "vendor": "DeepSeek", "date": "2026-10-15", "overall": 66.7, "categories": { "C1": 84, "C2": 70.8, "C3": 54, "C4": 60, "C5": 80, "C6": 53.3 }, "letters": 77.8, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "deepseek/deepseek-v4.1-flash", "vendor": "DeepSeek", "date": "2026-10-15", "overall": 62.3, "categories": { "C1": 84, "C2": 64.6, "C3": 44, "C4": 71.7, "C5": 88.9, "C6": 53.3 }, "letters": 60.7, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "google/gemini-3.8-flash", "vendor": "Google", "date": "2026-10-15", "overall": 73.5, "categories": { "C1": 96, "C2": 90, "C3": 66, "C4": 33.3, "C5": 97.8, "C6": 26.7 }, "letters": 56.5, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "google/gemini-4-argon", "vendor": "Google", "date": "2026-10-15", "overall": 74.1, "categories": { "C1": 100, "C2": 95, "C3": 64, "C4": 33.3, "C5": 93.3, "C6": 20 }, "letters": 52.3, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "meta/muse-wonder-1.0", "vendor": "Meta", "date": "2026-10-15", "overall": 84.6, "categories": { "C1": 96, "C2": 84.6, "C3": 68, "C4": 98.3, "C5": 88.9, "C6": 93.3 }, "letters": 74.7, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "mistral/mistral-large-4", "vendor": "Mistral", "date": "2026-10-15", "overall": 53.2, "categories": { "C1": 68, "C2": 76.3, "C3": 44, "C4": 8.3, "C5": 73.3, "C6": 13.3 }, "letters": 60.4, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "moonshotai/kimi-k3", "vendor": "Moonshot AI", "date": "2026-10-15", "overall": 60.2, "categories": { "C1": 64, "C2": 55.8, "C3": 40, "C4": 83.3, "C5": 80, "C6": 93.3 }, "letters": 70.2, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "openai/gpt-5.6-luna", "vendor": "OpenAI", "date": "2026-10-15", "overall": 72.9, "categories": { "C1": 92, "C2": 75, "C3": 48, "C4": 96.7, "C5": 100, "C6": 80 }, "letters": 83.2, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "openai/gpt-6-astra", "vendor": "OpenAI", "date": "2026-10-15", "overall": 84.9, "categories": { "C1": 88, "C2": 83.3, "C3": 74, "C4": 93.3, "C5": 93.3, "C6": 100 }, "letters": 92.1, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "openai/gpt-6-luna", "vendor": "OpenAI", "date": "2026-10-15", "overall": 80.8, "categories": { "C1": 96, "C2": 83.3, "C3": 56, "C4": 90, "C5": 100, "C6": 93.3 }, "letters": 91.1, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "openai/gpt-6.1-sol", "vendor": "OpenAI", "date": "2026-10-15", "overall": 83.7, "categories": { "C1": 92, "C2": 83.3, "C3": 66, "C4": 91.7, "C5": 93.3, "C6": 100 }, "letters": 92.1, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "spacexai/grok-4.7", "vendor": "xAI", "date": "2026-10-15", "overall": 66.5, "categories": { "C1": 60, "C2": 82.5, "C3": 44, "C4": 88.3, "C5": 93.3, "C6": 93.3 }, "letters": 78.8, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] }, { "model": "zai/glm-5.3", "vendor": "Zhipu (Z.ai)", "date": "2026-10-15", "overall": 60.3, "categories": { "C1": 56, "C2": 66.3, "C3": 40, "C4": 85, "C5": 77.8, "C6": 93.3 }, "letters": 60.2, "costUsd": null, "judges": [ "openai/gpt-6.1-sol", "google/gemini-4-argon" ] } ] }