Moondream: Moondream 3 Preview

{"about":"### Pratinjau Moondream 3\n\nMoondream 3 Preview termasuk keluarga Moondream yang berfokus pada pemahaman gambar dan tugas terkait objek tertentu dalam adegan. Pada generasi ketiga, pengembang beralih ke arsitektur campuran pakar: model memiliki total 9 miliar parameter dan menggunakan sekitar 2 miliar per token. Pembagian ini menggabungkan volume bobot keseluruhan dengan bagian aktif lebih kecil saat memproses setiap token.\n\nKartu resmi Preview menjelaskan 24 lapisan, encoder visual berbasis SigLIP, dan tokenizer SuperBPE. Keterampilan khusus model asal meliputi tanya jawab gambar, pembuatan keterangan, penunjukan titik, dan pembatasan objek dengan kotak. Ini tugas berbeda: penjelasan adegan, deskripsi seluruh gambar, dan penempatan spasial objek menghasilkan jenis keluaran berbeda.\n\n### Pekerjaan teks di Ailoot API\n\nUntuk model ini, Ailoot API menyediakan masukan dan jawaban teks. Gambar, koordinat objek, dan metode visual khusus tidak termasuk antarmuka saat ini. Namun, pengembang memberi contoh tersendiri untuk Preview berupa pertanyaan tanpa gambar: model menjelaskan konsep berdasarkan instruksi teks biasa.\n\nMode ini dapat digunakan untuk penjelasan, jawaban berdasarkan informasi yang diberikan, dan penataan materi teks. Misalnya, berikan draf deskripsi, tentukan audiens, dan minta penulisan ulang yang lebih jelas dengan mempertahankan fakta. Pilihan lain adalah melampirkan petunjuk teks dan meminta urutan tindakan dengan penjelasan singkat setiap langkah.\n\n### Merumuskan permintaan\n\nSebutkan tugas dan format jawaban secara langsung: penjelasan singkat, teks padu, daftar, atau tabel. Untuk menelaah materi tertentu, sertakan dalam pesan dan tentukan informasi yang harus tetap sama. Pertanyaan tentang adegan visual dalam mode teks memerlukan deskripsi lisan tertulis; model mengolah deskripsi yang diberikan, bukan gambar itu sendiri.\n\nAiloot API menyatakan konteks 32 000 token untuk Moondream 3 Preview, mencakup pesan dan materi diskusi saat ini. Penanda Preview menunjukkan rilis pendahuluan generasi ketiga; sifat versi ini dan versi 3.1 perlu dipertimbangkan terpisah, terutama kemampuan khusus model asal.","alias_of":null,"aliases":[],"author":"Moondream","canonical_model":"moondream/moondream3-preview","ctx":null,"formats":["chat","gemini","messages","responses"],"id":"moondream/moondream3-preview","input":["text"],"kind":"text","max_out":null,"methods":[{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"anthropic","fast_multiplier":1,"flex_multiplier":1,"fmt":"messages","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"Anthropic Messages","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1/messages","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"responses","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"OpenAI Responses","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1/responses","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"chat","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"OpenAI Chat","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1/chat/completions","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"gemini","fast_multiplier":1,"flex_multiplier":1,"fmt":"gemini","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"Gemini generateContent","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1beta/models/moondream/moondream3-preview:generateContent","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]}],"model_aliases":[{"canonical_model_id":"moondream/moondream3-preview","request_model_id":"moondream/moondream3-preview","route_label":"Ultra"}],"name":"Moondream: Moondream 3 Preview","price":{"in":0.6,"in_label":null,"out":5.0,"out_label":null,"sort_in":0.6,"sort_out":5.0,"varies":false},"public_id":"moondream/moondream3-preview","short":"Pratinjau Moondream 3 dengan arsitektur MoE ringkas; di Ailoot API menjawab pertanyaan dan instruksi teks."}