Voyage AI voyage-3 (2024年)（ボヤージュエーアイボヤージュ3）

概要

Voyage AI voyage-3 は、2024年9月Voyage AI発表voyage-3・Industry-leading retrieval-focused embedding model 2024年Voyage AI voyage-3 enterprise embedding position確立。voyage-3 specifications = Industry-leading retrieval-focused embedding (Industry-leading Voyage AI voyage-3 2024 + Industry-leading retrieval-focused embedding + Industry-leading Voyage AI flagship) + Industry-leading 1024d default + Industry-leading 32k context + Industry-leading $0.06/1M tokens Voyage AI voyage-3 enterprise embedding 2024。

主な特徴・仕組み

• Developer: Industry-leading Voyage AI (Founded 2023 Palo Alto California by Tengyu Ma + Xiang Li, ex-Stanford)
• Model: Industry-leading Voyage AI voyage-3 enterprise embedding model 2024年9月 release
• Retrieval-Focused: Industry-leading retrieval-focused signature (vs general-purpose)
• 1024d Default: Industry-leading 1024d default dimension signature
• 32k Context: Industry-leading 32k context window signature
• Quality Beats text-embedding-3-large: Industry-leading quality beats OpenAI text-embedding-3-large signature (per Voyage benchmarks)
• voyage-3-lite: Industry-leading voyage-3-lite lightweight variant
• voyage-3-large: Industry-leading voyage-3-large quality variant
• voyage-code-3 + voyage-finance-2 + voyage-law-2: Industry-leading domain-specific code/finance/law variants
• Voyage AI API: Industry-leading Voyage AI API + AWS Bedrock + AWS Marketplace
• Acquired by MongoDB: Industry-leading MongoDB acquisition 2025年2月 (rumor, post-2024 establishment)
• Voyage AI 2023 Palo Alto: Industry-leading Voyage AI 2023 Palo Alto Tengyu Ma + Xiang Li ex-Stanford 1 year
• Pricing: Industry-leading $0.06/1M tokens
• Industry-Leading: Industry-leading Voyage AI voyage-3 enterprise embedding 2024

スペック比較表

具体例・対応製品

• voyage-3 (2024年9月, $0.06/1M tokens, Voyage AI): Industry-leading Voyage AI voyage-3
• Industry-leading retrieval-focused vs general-purpose: Industry-leading retrieval-focused
• Industry-leading 1024d default + 32k context: Industry-leading 1024d + 32k context
• Industry-leading quality beats text-embedding-3-large: Industry-leading quality beats text-embedding-3-large
• Industry-leading voyage-3-lite + voyage-3-large + domain-specific variants: Industry-leading voyage-3 family
• 競合 OpenAI text-embedding-3-small + Cohere embed-v3 + BGE-M3 + Nomic Embed Text v1.5: Industry-leading embedding model competitors

選び方・注意点

Voyage AI voyage-3 は「Industry-leading retrieval-focused + 32k context」「Industry-leading quality + Voyage AI Stanford 1 year」用途のIndustry-leading premium 2024年9月 Voyage AI product。Industry-leading retrieval-focused (Industry-leading retrieval-focused signature vs general-purpose + Industry-leading voyage-3 retrieval signature + Industry-leading RAG-optimized embedding) で Industry-leading retrieval-focused + Industry-leading voyage-3 + Industry-leading RAG-optimized embedding。Industry-leading 32k context (Industry-leading 32k context window signature + Industry-leading voyage-3 32k signature + Industry-leading long-context embedding) で Industry-leading 32k context + Industry-leading voyage-3 + Industry-leading long-context embedding。Industry-leading quality beats text-embedding-3-large (Industry-leading quality beats OpenAI text-embedding-3-large signature per Voyage benchmarks + Industry-leading voyage-3 quality signature + Industry-leading state-of-art retrieval embedding) で Industry-leading quality beats text-embedding-3-large + Industry-leading voyage-3 quality + Industry-leading state-of-art retrieval embedding。Industry-leading domain-specific variants (Industry-leading voyage-code-3 + voyage-finance-2 + voyage-law-2 domain-specific + Industry-leading voyage-3 domain-specific signature + Industry-leading vertical embedding ecosystem) で Industry-leading domain-specific + Industry-leading voyage-3 domain + Industry-leading vertical embedding ecosystem。Industry-leading Voyage AI 2023 Palo Alto (Industry-leading Voyage AI 2023 Palo Alto Tengyu Ma + Xiang Li ex-Stanford 1 year + Industry-leading 1024d default + Industry-leading voyage-3-lite + voyage-3-large + Industry-leading Voyage AI API + AWS Bedrock + AWS Marketplace + Industry-leading $0.06/1M tokens) で Industry-leading Voyage AI 2023 + Industry-leading Palo Alto + Industry-leading Tengyu Ma + Xiang Li + Industry-leading ex-Stanford + Industry-leading 1 year + Industry-leading 1024d + Industry-leading voyage-3 family + Industry-leading Voyage AI API + AWS + Industry-leading $0.06/1M。但しIndustry-leading OpenAI text-embedding-3-small + Cohere embed-v3 + BGE-M3 + Nomic Embed Text v1.5 competition で Industry-leading 4-embedding model competitors vs Voyage AI voyage-3 + Industry-leading retrieval-focused vs general-purpose + 1024d default + 32k context + quality beats OpenAI text-embedding-3-large + voyage-3-lite + voyage-3-large + voyage-code-3 + voyage-finance-2 + voyage-law-2 domain-specific + Voyage AI API + AWS Bedrock + AWS Marketplace + Voyage AI 2023 Palo Alto Tengyu Ma + Xiang Li ex-Stanford 1 year + $0.06/1M tokens unique advantage adoption alignment必須。

関連用語との違い

• vs OpenAI text-embedding-3-small (2024 $0.02/1M): voyage-3は$0.06/1M + retrieval-focused + 32k Voyage AI Stanford・text-embedding-3-smallは$0.02/1M + Matryoshka + 8192 tokens OpenAI
• vs Cohere embed-v3 (2024 $0.10/1M): voyage-3は$0.06/1M + retrieval-focused + 32k + voyage family Voyage AI・embed-v3は$0.10/1M + 100+ languages + 3-variant Cohere
• vs BGE-M3 (2024 Free): voyage-3はAPI commercial $0.06/1M + retrieval-focused Voyage AI・BGE-M3はFree open-source + multi-functionality + 8192 tokens BAAI

よくある質問（FAQ）

Q1: Voyage AI voyage-3 vs OpenAI text-embedding-3-small 違いは？ A: voyage-3 (Industry-leading retrieval-focused + 1024d + 32k context + quality beats text-embedding-3-large per Voyage + voyage-3-lite + voyage-3-large + voyage-code-3 + voyage-finance-2 + voyage-law-2 + Voyage AI API + AWS + Voyage AI 2023 Palo Alto Tengyu Ma + Xiang Li ex-Stanford 1 year + $0.06/1M) vs text-embedding-3-small (Industry-leading 1536d default + Matryoshka adjustable 256-1536d + 5x cheaper $0.02/1M + 62.3 MTEB + 44.0 MIRACL + OpenAI API + Azure + replace ada-002 + multi-language + OpenAI 2015 SF Sam Altman + Greg Brockman + Ilya Sutskever 9 years)・Industry-leading retrieval-focused + 1024d + 32k context + quality beats + domain-specific + Voyage AI + Palo Alto + Tengyu Ma + Xiang Li + ex-Stanford + 2023 + 1 year + $0.06/1M = voyage-3 + Industry-leading 1536d + Matryoshka + 5x cheaper + 62.3 MTEB + 44.0 MIRACL + OpenAI + SF + Sam Altman + 9 years + $0.02/1M = text-embedding-3-small preference judgment (voyage-3 retrieval-focused + 32k + domain-specific Stanford vs text-embedding-3-small Matryoshka + 5x cheaper + multi-language OpenAI differ)。

Q2: Industry-leading retrieval-focused + 32k context value は？ A: Industry-leading retrieval-focused + 32k (Industry-leading retrieval-focused vs general-purpose + Industry-leading 32k context window + Industry-leading voyage-3 dual signature)。

Q3: Industry-leading quality + Voyage AI Stanford 1 year value は？ A: Industry-leading quality + Voyage AI Stanford (Industry-leading quality beats OpenAI text-embedding-3-large per Voyage benchmarks + Industry-leading Voyage AI 2023 Palo Alto Tengyu Ma + Xiang Li ex-Stanford 1 year heritage)。

まとめ

Voyage AI voyage-3 = 2024年9月Voyage AI発表のpremium enterprise embedding model。Industry-leading retrieval-focused signature vs general-purpose + Industry-leading 1024d default dimension signature + Industry-leading 32k context window signature + Industry-leading quality beats OpenAI text-embedding-3-large signature per Voyage benchmarks + Industry-leading voyage-3-lite lightweight variant + Industry-leading voyage-3-large quality variant + Industry-leading voyage-code-3 + voyage-finance-2 + voyage-law-2 domain-specific variants + Industry-leading Voyage AI API + AWS Bedrock + AWS Marketplace + Industry-leading Voyage AI 2023 Palo Alto California Tengyu Ma + Xiang Li ex-Stanford 1 year + Industry-leading $0.06/1M tokens + Industry-leading Voyage AI voyage-3 enterprise embedding 2024 position確立。