Cohere Embed v3
Cohere Embed v3 is a previous Cohere embedding snapshot, not the live catalog lead. Rechecked 2026-08-17: docs.cohere.com/docs/cohere-embed now leads with Embed 4. cohere.com/pricing did not print a public Embed v3 $0.10 / 1M row on the page we opened. Drop leftover top MTEB / 100+ languages as a 2026 census / $0.10 per 1M as the current default.
Compare text-embedding-3-large if you wanted a still-listed OpenAI v3 row, and BGE-M3 if you wanted a 2024 open multilingual card.
Key Features
- Previous SKU: Leftover copy described searchdocument / searchquery / classification / clustering and int8 compression. Recheck the live Embed docs before you quote those modes.
- Not the 2026 lead: Embed 4 is the current docs headline.
- No invented MTEB: 64.5+ stays off the table unless you re-run a first-party board.
- No leftover $0.10 row: Do not keep that dollar unless a live pricing cell still prints it.
Use Cases
- People with an Embed v3 bookmark who need the catalog corrected.
- People comparing leftover v3 vs live Embed 4.
- People who were about to paste $0.10 / 1M into a 2026 deck.
Limitation: We did not call Embed 4 or re-run MTEB.
Pricing
First-party pages on 2026-08-17.
| Piece | Price | Notes |
|---|---|---|
| Embed v3 | Previous snapshot | Not the live docs lead. Recheck pricing if you still have a v3 key. |
| Live Cohere embed | Recheck Embed 4 / pricing | Do not invent a $0.10 / 1M default. |
If a leftover blog still lists Embed v3 as the current Cohere default at $0.10 / 1M, treat the Embed 4 docs as the correction.
Getting Started
- Open docs.cohere.com/docs/cohere-embed before you hard-code
embed-english-v3.0. - Recheck Embed 4 if you wanted the live Cohere path.
- Recheck text-embedding-3-small if you wanted a public OpenAI dollar row.
- Do not paste MTEB 64.5 into a deck.
Frequently Asked Questions
Is Embed v3 still the Cohere default?
No. The docs we opened lead with Embed 4.
Is $0.10 / 1M still public?
Not on the pricing page we opened. Do not invent it.
Same as OpenAI ada-002?
No. ada-002 is an older billed OpenAI snapshot. This slug is leftover Cohere v3.
Alternatives
- text-embedding-3-large: Still-listed OpenAI v3 large row.
- text-embedding-3-small: Cheaper listed OpenAI v3 row.
- BGE-M3: 2024 open multilingual card.
Tips
- Call Embed v3 a previous snapshot.
- Point new Cohere work at Embed 4.
- Recheck pricing before you invent a v3 dollar.
Conclusion
Cohere Embed v3 is a previous snapshot, not the 2026 Cohere embedding default and not a public $0.10 / 1M row we could re-verify. Start at docs.cohere.com/docs/cohere-embed, then decide whether Embed 4 or an OpenAI v3 row already covers the path you need.
Comments
No comments yet. Be the first to comment!
Related Tools
Cohere Rerank 3.5
docs.cohere.com/docs/rerank
Previous Cohere rerank snapshot. Rechecked: live docs lead with rerank-v4.0. Not a $2/1k 2026 census.
Qwen3-Embedding
huggingface.co/Qwen
Previous Qwen3 embedding snapshot. Rechecked: still listed on Hugging Face. Drop leftover MTEB #1 as a 2026 census.
BGE-M3
huggingface.co/BAAI/bge-m3
2024 BAAI multilingual embedding snapshot. Rechecked: MIT card still listed. Not a 2026 MIRACL or MTEB crown.
Related Insights
Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.
A ChatGPT-signed Codex desktop app still shows mostly GPT in the picker. On Windows, enable only OpenCode Go and the Grok, GLM, Kimi, DeepSeek, and MiniMax models you already pay for appear in the same selector. Keys stay local. Native GPT stays put.
What Locks Codex Is the Picker, Not the Models
You already pay for OpenCode Go, Grok, and Z.ai, but the Codex picker still shows mostly GPT. The community project codex-router does not teach another install ritual. It puts subscriptions you already bought back into the selector, keeps keys on the machine, and leaves native GPT alone.
Seven AI Coding CLIs, Six Months: No Matter How Strong the Model, Work Needs Supervision
Claude Code, Codex, opencode, pi, omp and DeepSeek Harness all have personalities. After six months of deep use I run a division of labor: pi for the fastest cheapest reviews, omp for complex PRs, DeepSeek Harness on V4 Flash for high-frequency low-cost review, and Claude Code, Qoder and Cursor for writing. No matter how strong the model, work needs supervision — ideally from an independent third party.