text-embedding-3-large ranks #3 of 12 in our embedding models testing. the default everyone reaches for, and it is showing its age.
83/100
still a solid general-purpose embedding with the most flexible dimension control here — on an 8k context and a model that has not been updated since january 2024.
why people look for an alternative
−8,192-token context is joint smallest in the ranking
−no replacement shipped since january 2024
−no multilingual claim, no multimodal support, no reranker sibling
−quotes an MTEB score without naming the board version
stay with text-embedding-3-large if arbitrary dimension truncation via the dimensions parameter, not a fixed list is the thing you care about most — nothing below beats it on that.
#1 in embedding models · apache-2.0, best open multilingual quality, and a tenth of a cent hosted
89/100
verdictthe best combination of open licence, multilingual quality and truncatable dimensions in the category — provided you can afford to serve an 8b model or are happy renting one.
Qwen3-Embedding-8B vs text-embedding-3-large
text-embedding-3-large
Qwen3-Embedding-8B
price
$0.13 / 1m tokens
$0.010 / 1m tokens on DeepInfra
free tier
no
yes
price / 1m
$0.13
$0.010 hosted, free self-hosted
dimensions
3,072, truncatable to any size
4,096, truncatable to 32
context
8,192 tokens
32,768 tokens
licence
proprietary api
Apache-2.0
multimodal
no — text only
no — text only
switch formultilingual retrieval where the licence has to be clean and the bill has to be small
pros
+apache-2.0 on code and weights — commercially safe without conditions
+32,768-token context, four times OpenAI's or Gemini's
+matryoshka truncation from 4,096 down to 32 dimensions
+publishes MTEB scores with the board version named
+matching apache-2.0 reranker family
cons
−8b parameters — the heaviest model here to self-host
−4,096 native dimensions is a large index if you do not truncate
#2 in embedding models · one vector space for text, images, audio and video
86/100
verdictthe only model here that puts text, images, audio and video in one shared space — at the highest text price in the ranking, and it just invalidated its own predecessor's vectors.
gemini-embedding-2 vs text-embedding-3-large
text-embedding-3-large
gemini-embedding-2
price
$0.13 / 1m tokens
$0.20 / 1m text tokens
free tier
no
yes
price / 1m
$0.13
$0.20 text, $0.10 batch
dimensions
3,072, truncatable to any size
128 to 3,072
context
8,192 tokens
8,192 tokens
licence
proprietary api
proprietary api
multimodal
no — text only
yes — text, image, audio, video
switch forretrieval across mixed media where text-only embeddings cannot answer the question
pros
+text, images, audio, video and documents in one shared vector space
+100+ languages, with matryoshka dimensions from 128 to 3,072
+batch mode halves the text rate to $0.10 per million
+8,192-token context, four times its own predecessor
cons
−$0.20 per million is the highest text rate in the ranking
−audio costs 32x and video 60x the text rate
−vectors are incompatible with gemini-embedding-001 — upgrading means re-embedding everything
#5 in embedding models · 128k of context, text and pdfs in one space, and no published price
78/100
verdictby far the longest context here and genuine text-plus-pdf embedding — from the only vendor in this ranking that will not tell you what an api call costs.
Cohere Embed v4 vs text-embedding-3-large
text-embedding-3-large
Cohere Embed v4
price
$0.13 / 1m tokens
not published per token
free tier
no
yes
price / 1m
$0.13
not published
dimensions
3,072, truncatable to any size
256 / 512 / 1,024 / 1,536
context
8,192 tokens
128,000 tokens
licence
proprietary api
proprietary api
multimodal
no — text only
yes — text, images, pdfs
switch forlong-document and pdf retrieval where a 128k window removes the chunking problem
pros
+128,000-token context — fifteen times OpenAI's, the longest here by far
+text, images and pdfs embedded into one space
+matching Rerank 3.5 and Rerank 4 models
+four selectable output dimensions from 256 to 1,536
cons
−no per-token api price published anywhere on cohere.com
−only dedicated-instance pricing is public, from $2,500 a month
−no MTEB score on its own pages, and third-party figures disagree
#10 in embedding models · a flat price, a tidy api, and almost no published specification
62/100
verdictcheaper than OpenAI and easy to adopt if you are already a Mistral customer — from a model that publishes neither its dimensions nor its benchmark scores.
Mistral Embed vs text-embedding-3-large
text-embedding-3-large
Mistral Embed
price
$0.13 / 1m tokens
$0.10 / 1m tokens
free tier
no
yes
price / 1m
$0.13
$0.10
dimensions
3,072, truncatable to any size
not published by vendor
context
8,192 tokens
8,000 tokens
licence
proprietary api
proprietary api
multimodal
no — text only
no — text and code
switch forteams already building on Mistral who want embeddings on the same bill
pros
+$0.10 per million, cheaper than OpenAI's flagship
+no new vendor if you are already using the Mistral api
+flat pricing with no tiers or dimension-based surcharges
+a separate Codestral Embed exists for code retrieval
cons
−output dimensions are not published by the vendor
−no parameter count, language count or MTEB score published
−dates to december 2023 — the oldest model in this ranking
−no weights, no self-host path, no reranker sibling
#11 in embedding models · excellent quality per parameter, on weights you may not use commercially
58/100
verdicta genuinely strong sub-1b multilingual model that most readers of this page cannot legally deploy — and the licence is not what its reputation suggests.
Jina Embeddings v5 vs text-embedding-3-large
text-embedding-3-large
Jina Embeddings v5
price
$0.13 / 1m tokens
not published per token
free tier
no
yes
price / 1m
$0.13
not published
dimensions
3,072, truncatable to any size
1,024, truncatable to 32
context
8,192 tokens
32,768 tokens
licence
proprietary api
CC-BY-NC-4.0 — non-commercial
multimodal
no — text only
no — v5-omni is a separate model
switch forresearch and evaluation work where the non-commercial terms are not a problem
pros
+677m parameters with 32,768 tokens of context
+matryoshka truncation from 1,024 down to 32 dimensions
+93 languages claimed from a sub-billion-parameter model
+a separate v5-omni line covers images, audio, video and pdfs
cons
−weights are CC-BY-NC-4.0 — commercial use prohibited without a separate licence
−v4 was non-commercial too, under a qwen research licence
−no per-token price published for the hosted api
−the v4-to-v5 split moved multimodality to a different model line
every tool on this page went through the same test as text-embedding-3-large — same tasks, same order, scored the same way. the comparison tables are the figures from that testing, not vendor spec sheets.