{"id":121623,"date":"2026-10-02T02:55:51","date_gmt":"2026-10-02T02:55:51","guid":{"rendered":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/"},"modified":"2026-10-02T02:55:51","modified_gmt":"2026-10-02T02:55:51","slug":"cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai","status":"publish","type":"post","link":"https:\/\/youzum.net\/zh\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/","title":{"rendered":"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI"},"content":{"rendered":"<p class=\"wp-block-paragraph\">Cohere has released <a href=\"https:\/\/cohere.com\/blog\/embed-5\">Embed 5<\/a>, a new embedding model family. It targets enterprise search, RAG, and agentic retrieval. The model family ships in 2 tiers. <strong>Embed 5 Pro<\/strong> targets maximum retrieval quality. <strong>Embed 5 Fast<\/strong> targets latency and cost on the live query path. Both accept text, images, and fused text plus image inputs. Both cover 100+ languages and read up to 128K tokens. The key design choice: Pro and Fast share 1 embedding space. You can index with one and query with the other.<\/p>\n<p class=\"wp-block-paragraph\"><strong>Is it deployable today?<\/strong> Yes, both tiers are generally available on the <a href=\"https:\/\/dashboard.cohere.com\/\">Cohere API and Model Vault<\/a>, <a href=\"https:\/\/ai.azure.com\/catalog\/models\/Cohere-Embed-V5-Pro\">Microsoft Foundry<\/a>, and <a href=\"https:\/\/aws.amazon.com\/marketplace\/pp\/prodview-pobc74bcuaohi\">Amazon SageMaker<\/a>. Private VPC or on-prem serving runs through vLLM.<\/p>\n<h2 class=\"wp-block-heading\"><strong>What Cohere Shipped<\/strong><\/h2>\n<p class=\"wp-block-paragraph\">The API model IDs are <code>embed-v5.0-pro<\/code> and <code>embed-v5.0-fast<\/code>, per <a href=\"https:\/\/docs.cohere.com\/docs\/cohere-embed\">Cohere\u2019s model docs<\/a>. Both output 2048, 1536, 1024, 768, 512, or 256 dimensions, with 2048 as default. Embeddings come back as float, int8, or binary. Pro costs $0.12 per 1M text tokens. Fast costs $0.08. Image inputs cost $0.40 per 1M tokens on both.<\/p>\n<p class=\"wp-block-paragraph\">Embed 5 can embed a page image directly. It can also fuse an image with its metadata into a single vector. That is important for scanned pages, slide decks, schematics, and charts, where text extraction drops information.<\/p>\n<h2 class=\"wp-block-heading\"><strong>Pro and Fast: One Index, Two Query Paths<\/strong><\/h2>\n<p class=\"wp-block-paragraph\">Cohere tested every corpus and query pairing across 40 development datasets. Normalized to Pro plus Pro at 100, a Pro index queried with Fast scored 98.4. An all-Fast setup scored 96.6. Cohere\u2019s recommended pattern is to index with Pro and query with Fast. One constraint: both sides must use the same output dimension.<\/p>\n<p class=\"wp-block-paragraph\">The split targets agentic workloads. An agent may issue dozens of searches per task, and query latency compounds. Cohere team reports Fast processed 377.3 documents per second versus 159.7 for Pro.<\/p>\n<h2 class=\"wp-block-heading\"><strong>Benchmarks<\/strong><\/h2>\n<p class=\"wp-block-paragraph\">On <a href=\"https:\/\/huggingface.co\/blog\/QuentinJG\/introducing-vidore-v3\">ViDoRe V3<\/a>, Embed 5 Pro averages 85.8, an 8.8-point gain over Embed 4. Fast averages 84.5. Voyage 4 Large scores 83.7, Gemini Embedding 2 scores 83.2, and OpenAI text-embedding-3-large scores 75.5. On Cohere\u2019s parsed-PDF suite, Pro leads at 84.8 against Voyage 4 Large at 83.6.<\/p>\n<p class=\"wp-block-paragraph\">Finance is the strongest showing. Pro ranks first on <a href=\"https:\/\/arxiv.org\/abs\/2311.11944\">FinanceBench<\/a> (80.1), <a href=\"https:\/\/arxiv.org\/abs\/2109.00122\">FinQA<\/a> (90.0), and ViDoRe V3 Finance (85.0). Fast ranks second on all 3.<\/p>\n<p class=\"wp-block-paragraph\">Multilingual results are mixed. Pro leads the European-language average at 77. However, Gemini Embedding 2 beats Pro on 9 of 10 further languages in <a href=\"https:\/\/docs.google.com\/spreadsheets\/d\/1HLj1KOjJdFh9bTAc6lbWxG7VfjkRMeE7V11Q0eojtAA\/edit?gid=0#gid=0\">Cohere\u2019s own results table<\/a>. Those include Japanese, Arabic, Hindi, and Telugu.<\/p>\n<p class=\"wp-block-paragraph\"><strong>One important thing to note<\/strong>. Most numbers use <a href=\"https:\/\/cohere.com\/blog\/rcp-ndcg\">RCP-nDCG@10<\/a>, a new Cohere metric. It reorders a fixed candidate set, so it measures reranking quality more than first-stage retrieval. Cohere published the <a href=\"https:\/\/github.com\/cohere-ai\/rcp-ndcg\">evaluation code<\/a>, but independent replication is still pending.<\/p>\n<h2 class=\"wp-block-heading\"><strong>Storage Costs at Scale<\/strong><\/h2>\n<p class=\"wp-block-paragraph\">Embed 5 uses Matryoshka representation learning plus lower-precision outputs. A 2048-dim float32 vector needs 8 KB. A 1024-dim int8 vector needs 1 KB. A 256-dim binary vector needs 32 bytes. Across 100M chunks, raw storage drops from about 819 GB to 3.2 GB. Cohere recommends 1024-dim int8 as the default, citing near-full-precision quality.<\/p>\n<h2 class=\"wp-block-heading\"><strong>Interactive Explainer: How Embed 5 Works<\/strong><\/h2>\n<p><title>Cohere Embed 5 Explainer | Marktechpost<\/title><\/p>\n<div>\n<div class=\"hd\"><span class=\"tag\">Cohere Embed 5 \u00b7 Interactive explainer<\/span><span class=\"cnt\">1 \/ 6<\/span><\/div>\n<div class=\"bar-top\"><i><\/i><\/div>\n<p><!-- 1 --><\/p>\n<section class=\"slide on\" data-i=\"0\">\n<h3>1. What an embedding model actually does<\/h3>\n<p class=\"sub\">Text goes in, a vector of numbers comes out. Search then ranks documents by how close their vectors sit to the query vector. Press the button to watch it run.<\/p>\n<div class=\"card\">\n<p class=\"lbl\">Query<\/p>\n<div class=\"q\">What happened to net interest margin last quarter?<\/div>\n<div class=\"row\"><button class=\"btn go pulse\">Embed and search<\/button><span class=\"note\">Model: embed-v5.0-pro \u00b7 1024 dims<\/span><\/div>\n<div class=\"vec\"><\/div>\n<p class=\"lbl\">Documents ranked by cosine similarity<\/p>\n<div>\n<div class=\"doc\" data-s=\"0.71\"><span class=\"t\">Net interest margin narrowed 12 bps to 2.61% as deposit costs rose.<\/span><span class=\"m\"><i><\/i><\/span><span class=\"v\">0.00<\/span><\/div>\n<div class=\"doc\" data-s=\"0.18\"><span class=\"t\">Torque the mounting bolts to 45 Nm in a star pattern before refitting the cover.<\/span><span class=\"m\"><i><\/i><\/span><span class=\"v\">0.00<\/span><\/div>\n<div class=\"doc\" data-s=\"0.22\"><span class=\"t\">Employees accrue 1.5 days of paid leave for each month of service.<\/span><span class=\"m\"><i><\/i><\/span><span class=\"v\">0.00<\/span><\/div>\n<\/div>\n<p class=\"note\">Documents come from Cohere\u2019s launch code sample. Vector bars and similarity scores here are illustrative, not real model output.<\/p>\n<\/div>\n<\/section>\n<p><!-- 2 --><\/p>\n<section class=\"slide\" data-i=\"1\">\n<h3>2. Two tiers, one embedding space<\/h3>\n<p class=\"sub\">Pro and Fast write vectors into the same space. Pick which model indexes the corpus and which one embeds the query. Quality is normalized so Pro plus Pro equals 100.<\/p>\n<div class=\"grid2\">\n<div class=\"card\">\n<p class=\"lbl\">Index corpus with<\/p>\n<div class=\"row\"><button class=\"btn act\" data-c=\"pro\">Pro<\/button><button class=\"btn\" data-c=\"fast\">Fast<\/button><\/div>\n<p class=\"lbl\">Embed queries with<\/p>\n<div class=\"row\"><button class=\"btn\" data-q=\"pro\">Pro<\/button><button class=\"btn act\" data-q=\"fast\">Fast<\/button><\/div>\n<p class=\"note\">Cohere\u2019s recommended pattern: index with Pro, query with Fast.<\/p>\n<\/div>\n<div class=\"card\">\n<div class=\"gauge\"><svg width=\"170\" height=\"170\" viewbox=\"0 0 170 170\"><\/svg>\n<div class=\"val\"><b>0<\/b><small>relative nDCG@10<\/small><\/div>\n<\/div>\n<p class=\"note\">Ring zoomed to the 90 to 100 range so small gaps are visible.<\/p>\n<\/div>\n<\/div>\n<div class=\"card\">\n<p class=\"lbl\">Document throughput (docs per second, Cohere-measured)<\/p>\n<div class=\"hb\">\n<div class=\"n\"><span>Embed 5 Fast<\/span><em>377.3<\/em><\/div>\n<div class=\"tr\"><i data-w=\"100\"><\/i><\/div>\n<\/div>\n<div class=\"hb coral\">\n<div class=\"n\"><span>Embed 5 Pro<\/span><em>159.7<\/em><\/div>\n<div class=\"tr\"><i data-w=\"42.3\"><\/i><\/div>\n<\/div>\n<p class=\"note\">Price per 1M text tokens: Pro $0.12, Fast $0.08. Images cost $0.40 per 1M tokens on both tiers.<\/p>\n<\/div>\n<\/section>\n<p><!-- 3 --><\/p>\n<section class=\"slide\" data-i=\"2\">\n<h3>3. Shrink the vector, not the quality<\/h3>\n<p class=\"sub\">Matryoshka training lets you truncate dimensions. Lower precision outputs cut bytes further. Change the settings and watch storage move.<\/p>\n<div class=\"card\">\n<p class=\"lbl\">Output dimensions<\/p>\n<div class=\"row\"><button class=\"btn act\" data-d=\"2048\">2048<\/button><button class=\"btn\" data-d=\"1536\">1536<\/button><button class=\"btn\" data-d=\"1024\">1024<\/button><button class=\"btn\" data-d=\"768\">768<\/button><button class=\"btn\" data-d=\"512\">512<\/button><button class=\"btn\" data-d=\"256\">256<\/button><\/div>\n<p class=\"lbl\">Precision<\/p>\n<div class=\"row\"><button class=\"btn act\" data-p=\"4\">float32<\/button><button class=\"btn\" data-p=\"1\">int8<\/button><button class=\"btn\" data-p=\"0.125\">binary<\/button><\/div>\n<p class=\"lbl\">Chunks in your index: <span>100,000,000<\/span><\/p>\n<div class=\"kv\">\n<div><b>8,192 B<\/b><small>bytes per vector<\/small><\/div>\n<div><b>819.2 GB<\/b><small>raw vector storage<\/small><\/div>\n<div><b>1x<\/b><small>smaller than 2048 float32<\/small><\/div>\n<\/div>\n<div class=\"hb\">\n<div class=\"n\"><span>Storage vs 2048-dim float32 baseline<\/span><em>100%<\/em><\/div>\n<div class=\"tr\"><i><\/i><\/div>\n<\/div>\n<p class=\"note\">Math: dims x bytes per value (4, 1, or 1\/8). Cohere recommends 1024-dim int8 as the default sweet spot. Binary trades some accuracy and suits a first pass before reranking.<\/p>\n<\/div>\n<\/section>\n<p><!-- 4 --><\/p>\n<section class=\"slide\" data-i=\"3\">\n<h3>4. How it scores against rivals<\/h3>\n<p class=\"sub\">Cohere-reported averages using its new RCP-nDCG@10 metric. Treat these as vendor numbers until independent runs land.<\/p>\n<div class=\"row\"><button class=\"btn act\" data-t=\"v\">ViDoRe V3<\/button><button class=\"btn\" data-t=\"p\">Parsed PDFs<\/button><button class=\"btn\" data-t=\"f\">Finance<\/button><\/div>\n<div class=\"card\"><\/div>\n<p class=\"note\">Source: Cohere Embed 5 launch post, Sep 30 2026. RCP-nDCG@10 reorders a fixed candidate set, so it reflects reranking quality more than first-stage recall.<\/p>\n<\/section>\n<p><!-- 5 --><\/p>\n<section class=\"slide\" data-i=\"4\">\n<h3>5. How much fits in one input<\/h3>\n<p class=\"sub\">A longer context means fewer chunks for long filings and manuals. Press play to compare maximum input length per call.<\/p>\n<div class=\"card\">\n<button class=\"btn go\">Compare context windows<\/button>\n<div class=\"hb\">\n<div class=\"n\"><span>Cohere Embed 5 Pro \/ Fast<\/span><em>128K tokens<\/em><\/div>\n<div class=\"tr\"><i data-w=\"100\"><\/i><\/div>\n<\/div>\n<div class=\"hb other\">\n<div class=\"n\"><span>Voyage 4 Large<\/span><em>32,000 tokens<\/em><\/div>\n<div class=\"tr\"><i data-w=\"24.4\"><\/i><\/div>\n<\/div>\n<div class=\"hb other\">\n<div class=\"n\"><span>Gemini Embedding 2<\/span><em>8,192 tokens<\/em><\/div>\n<div class=\"tr\"><i data-w=\"6.25\"><\/i><\/div>\n<\/div>\n<div class=\"hb other\">\n<div class=\"n\"><span>OpenAI text-embedding-3-large<\/span><em>8,191 tokens<\/em><\/div>\n<div class=\"tr\"><i data-w=\"6.25\"><\/i><\/div>\n<\/div>\n<p class=\"note\">Limits taken from each vendor\u2019s model docs. 128K is shown as 131,072 tokens for bar scale.<\/p>\n<\/div>\n<\/section>\n<p><!-- 6 --><\/p>\n<section class=\"slide\" data-i=\"5\">\n<h3>6. Is it deployable? Yes. Pick a route.<\/h3>\n<p class=\"sub\">Both tiers are generally available. Tap a route to see what it gives you.<\/p>\n<div class=\"dep\">\n<button class=\"act\" data-k=\"0\"><b>Cohere API<\/b><small>Managed endpoint<\/small><\/button><br \/>\n<button data-k=\"1\"><b>Model Vault<\/b><small>Dedicated inference<\/small><\/button><br \/>\n<button data-k=\"2\"><b>Microsoft Foundry<\/b><small>Azure catalog<\/small><\/button><br \/>\n<button data-k=\"3\"><b>Amazon SageMaker<\/b><small>AWS Marketplace<\/small><\/button><br \/>\n<button data-k=\"4\"><b>Private VPC \/ on-prem<\/b><small>Served with vLLM<\/small><\/button><br \/>\n<button data-k=\"5\"><b>Cohere North<\/b><small>Built-in retrieval<\/small><\/button>\n<\/div>\n<div class=\"card detail\"><\/div>\n<\/section>\n<div class=\"nav\">\n<button class=\"btn\">\u2190 Back<\/button>\n<div class=\"dots\"><\/div>\n<p><button class=\"btn act\">Next \u2192<\/button>\n<\/p><\/div>\n<div class=\"ft\"><a href=\"https:\/\/www.marktechpost.com\/\" target=\"_blank\" rel=\"noopener\">Built by Marktechpost<\/a><span class=\"src\">Data: <a href=\"https:\/\/cohere.com\/blog\/embed-5\" target=\"_blank\" rel=\"noopener\">Cohere blog<\/a> \u00b7 <a href=\"https:\/\/docs.cohere.com\/docs\/cohere-embed\" target=\"_blank\" rel=\"noopener\">Cohere docs<\/a><\/span><\/div>\n<\/div>\n<div class=\"tr\"><i data-w=\"'+((r[1]-lo)\/(100-lo)*100).toFixed(1)+'\"><\/i><\/div>","protected":false},"excerpt":{"rendered":"<p>Cohere has released Embed 5, a new embedding model family. It targets enterprise search, RAG, and agentic retrieval. The model family ships in 2 tiers. Embed 5 Pro targets maximum retrieval quality. Embed 5 Fast targets latency and cost on the live query path. Both accept text, images, and fused text plus image inputs. Both cover 100+ languages and read up to 128K tokens. The key design choice: Pro and Fast share 1 embedding space. You can index with one and query with the other. Is it deployable today? Yes, both tiers are generally available on the Cohere API and Model Vault, Microsoft Foundry, and Amazon SageMaker. Private VPC or on-prem serving runs through vLLM. What Cohere Shipped The API model IDs are embed-v5.0-pro and embed-v5.0-fast, per Cohere\u2019s model docs. Both output 2048, 1536, 1024, 768, 512, or 256 dimensions, with 2048 as default. Embeddings come back as float, int8, or binary. Pro costs $0.12 per 1M text tokens. Fast costs $0.08. Image inputs cost $0.40 per 1M tokens on both. Embed 5 can embed a page image directly. It can also fuse an image with its metadata into a single vector. That is important for scanned pages, slide decks, schematics, and charts, where text extraction drops information. Pro and Fast: One Index, Two Query Paths Cohere tested every corpus and query pairing across 40 development datasets. Normalized to Pro plus Pro at 100, a Pro index queried with Fast scored 98.4. An all-Fast setup scored 96.6. Cohere\u2019s recommended pattern is to index with Pro and query with Fast. One constraint: both sides must use the same output dimension. The split targets agentic workloads. An agent may issue dozens of searches per task, and query latency compounds. Cohere team reports Fast processed 377.3 documents per second versus 159.7 for Pro. Benchmarks On ViDoRe V3, Embed 5 Pro averages 85.8, an 8.8-point gain over Embed 4. Fast averages 84.5. Voyage 4 Large scores 83.7, Gemini Embedding 2 scores 83.2, and OpenAI text-embedding-3-large scores 75.5. On Cohere\u2019s parsed-PDF suite, Pro leads at 84.8 against Voyage 4 Large at 83.6. Finance is the strongest showing. Pro ranks first on FinanceBench (80.1), FinQA (90.0), and ViDoRe V3 Finance (85.0). Fast ranks second on all 3. Multilingual results are mixed. Pro leads the European-language average at 77. However, Gemini Embedding 2 beats Pro on 9 of 10 further languages in Cohere\u2019s own results table. Those include Japanese, Arabic, Hindi, and Telugu. One important thing to note. Most numbers use RCP-nDCG@10, a new Cohere metric. It reorders a fixed candidate set, so it measures reranking quality more than first-stage retrieval. Cohere published the evaluation code, but independent replication is still pending. Storage Costs at Scale Embed 5 uses Matryoshka representation learning plus lower-precision outputs. A 2048-dim float32 vector needs 8 KB. A 1024-dim int8 vector needs 1 KB. A 256-dim binary vector needs 32 bytes. Across 100M chunks, raw storage drops from about 819 GB to 3.2 GB. Cohere recommends 1024-dim int8 as the default, citing near-full-precision quality. Interactive Explainer: How Embed 5 Works Cohere Embed 5 Explainer | Marktechpost Cohere Embed 5 \u00b7 Interactive explainer1 \/ 6 1. What an embedding model actually does Text goes in, a vector of numbers comes out. Search then ranks documents by how close their vectors sit to the query vector. Press the button to watch it run. Query What happened to net interest margin last quarter? Embed and searchModel: embed-v5.0-pro \u00b7 1024 dims Documents ranked by cosine similarity Net interest margin narrowed 12 bps to 2.61% as deposit costs rose.0.00 Torque the mounting bolts to 45 Nm in a star pattern before refitting the cover.0.00 Employees accrue 1.5 days of paid leave for each month of service.0.00 Documents come from Cohere\u2019s launch code sample. Vector bars and similarity scores here are illustrative, not real model output. 2. Two tiers, one embedding space Pro and Fast write vectors into the same space. Pick which model indexes the corpus and which one embeds the query. Quality is normalized so Pro plus Pro equals 100. Index corpus with ProFast Embed queries with ProFast Cohere\u2019s recommended pattern: index with Pro, query with Fast. 0relative nDCG@10 Ring zoomed to the 90 to 100 range so small gaps are visible. Document throughput (docs per second, Cohere-measured) Embed 5 Fast377.3 Embed 5 Pro159.7 Price per 1M text tokens: Pro $0.12, Fast $0.08. Images cost $0.40 per 1M tokens on both tiers. 3. Shrink the vector, not the quality Matryoshka training lets you truncate dimensions. Lower precision outputs cut bytes further. Change the settings and watch storage move. Output dimensions 204815361024768512256 Precision float32int8binary Chunks in your index: 100,000,000 8,192 Bbytes per vector 819.2 GBraw vector storage 1xsmaller than 2048 float32 Storage vs 2048-dim float32 baseline100% Math: dims x bytes per value (4, 1, or 1\/8). Cohere recommends 1024-dim int8 as the default sweet spot. Binary trades some accuracy and suits a first pass before reranking. 4. How it scores against rivals Cohere-reported averages using its new RCP-nDCG@10 metric. Treat these as vendor numbers until independent runs land. ViDoRe V3Parsed PDFsFinance Source: Cohere Embed 5 launch post, Sep 30 2026. RCP-nDCG@10 reorders a fixed candidate set, so it reflects reranking quality more than first-stage recall. 5. How much fits in one input A longer context means fewer chunks for long filings and manuals. Press play to compare maximum input length per call. Compare context windows Cohere Embed 5 Pro \/ Fast128K tokens Voyage 4 Large32,000 tokens Gemini Embedding 28,192 tokens OpenAI text-embedding-3-large8,191 tokens Limits taken from each vendor\u2019s model docs. 128K is shown as 131,072 tokens for bar scale. 6. Is it deployable? Yes. Pick a route. Both tiers are generally available. Tap a route to see what it gives you. Cohere APIManaged endpoint Model VaultDedicated inference Microsoft FoundryAzure catalog Amazon SageMakerAWS Marketplace Private VPC \/ on-premServed with vLLM Cohere NorthBuilt-in retrieval \u2190 Back Next \u2192 Built by MarktechpostData: Cohere blog \u00b7 Cohere docs<\/p>","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"pmpro_default_level":"","site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"_pvb_checkbox_block_on_post":false,"footnotes":""},"categories":[52,5,7,1],"tags":[],"class_list":["post-121623","post","type-post","status-publish","format-standard","hentry","category-ai-club","category-committee","category-news","category-uncategorized","pmpro-has-access"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v25.3 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI - YouZum<\/title>\n<meta name=\"description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/youzum.net\/zh\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI - YouZum\" \/>\n<meta property=\"og:description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta property=\"og:url\" content=\"https:\/\/youzum.net\/zh\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/\" \/>\n<meta property=\"og:site_name\" content=\"YouZum\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/DroneAssociationTH\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-02T02:55:51+00:00\" \/>\n<meta name=\"author\" content=\"admin NU\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin NU\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/\"},\"author\":{\"name\":\"admin NU\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\"},\"headline\":\"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI\",\"datePublished\":\"2026-10-02T02:55:51+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/\"},\"wordCount\":930,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"articleSection\":[\"AI\",\"Committee\",\"News\",\"Uncategorized\"],\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/\",\"url\":\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/\",\"name\":\"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI - YouZum\",\"isPartOf\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#website\"},\"datePublished\":\"2026-10-02T02:55:51+00:00\",\"description\":\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\",\"breadcrumb\":{\"@id\":\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/youzum.net\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/yousum.gpucore.co\/#website\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"name\":\"YouSum\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/yousum.gpucore.co\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\",\"name\":\"Drone Association Thailand\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"width\":300,\"height\":300,\"caption\":\"Drone Association Thailand\"},\"image\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/DroneAssociationTH\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\",\"name\":\"admin NU\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"caption\":\"admin NU\"},\"url\":\"https:\/\/youzum.net\/zh\/members\/adminnu\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI - YouZum","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/youzum.net\/zh\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/","og_locale":"zh_CN","og_type":"article","og_title":"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI - YouZum","og_description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","og_url":"https:\/\/youzum.net\/zh\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/","og_site_name":"YouZum","article_publisher":"https:\/\/www.facebook.com\/DroneAssociationTH\/","article_published_time":"2026-10-02T02:55:51+00:00","author":"admin NU","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin NU","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"5 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#article","isPartOf":{"@id":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/"},"author":{"name":"admin NU","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c"},"headline":"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI","datePublished":"2026-10-02T02:55:51+00:00","mainEntityOfPage":{"@id":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/"},"wordCount":930,"commentCount":0,"publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"articleSection":["AI","Committee","News","Uncategorized"],"inLanguage":"zh-Hans","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/","url":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/","name":"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI - YouZum","isPartOf":{"@id":"https:\/\/yousum.gpucore.co\/#website"},"datePublished":"2026-10-02T02:55:51+00:00","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","breadcrumb":{"@id":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/youzum.net\/cohere-releases-embed-5-how-it-compares-to-voyage-4-large-gemini-embedding-2-and-openai\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/youzum.net\/"},{"@type":"ListItem","position":2,"name":"Cohere Releases Embed 5: How It Compares to Voyage 4 Large, Gemini Embedding 2, and OpenAI"}]},{"@type":"WebSite","@id":"https:\/\/yousum.gpucore.co\/#website","url":"https:\/\/yousum.gpucore.co\/","name":"YouSum","description":"","publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/yousum.gpucore.co\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"zh-Hans"},{"@type":"Organization","@id":"https:\/\/yousum.gpucore.co\/#organization","name":"Drone Association Thailand","url":"https:\/\/yousum.gpucore.co\/","logo":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","width":300,"height":300,"caption":"Drone Association Thailand"},"image":{"@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/DroneAssociationTH\/"]},{"@type":"Person","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c","name":"admin NU","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","caption":"admin NU"},"url":"https:\/\/youzum.net\/zh\/members\/adminnu\/"}]}},"rttpg_featured_image_url":null,"rttpg_author":{"display_name":"admin NU","author_link":"https:\/\/youzum.net\/zh\/members\/adminnu\/"},"rttpg_comment":0,"rttpg_category":"<a href=\"https:\/\/youzum.net\/zh\/category\/ai-club\/\" rel=\"category tag\">AI<\/a> <a href=\"https:\/\/youzum.net\/zh\/category\/committee\/\" rel=\"category tag\">Committee<\/a> <a href=\"https:\/\/youzum.net\/zh\/category\/news\/\" rel=\"category tag\">News<\/a> <a href=\"https:\/\/youzum.net\/zh\/category\/uncategorized\/\" rel=\"category tag\">Uncategorized<\/a>","rttpg_excerpt":"Cohere has released Embed 5, a new embedding model family. It targets enterprise search, RAG, and agentic retrieval. The model family ships in 2 tiers. Embed 5 Pro targets maximum retrieval quality. Embed 5 Fast targets latency and cost on the live query path. Both accept text, images, and fused text plus image inputs. Both&hellip;","_links":{"self":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/posts\/121623","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/comments?post=121623"}],"version-history":[{"count":0,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/posts\/121623\/revisions"}],"wp:attachment":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/media?parent=121623"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/categories?post=121623"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/tags?post=121623"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}