BGE-M3
BAAI
BAAI's BGE-M3 multilingual text embedding model with an 8K-token input limit and 1024-dimension vectors.
Embeddingsus-eu-hosted
baai/bge-m3
Also callable as bge-m3
Call it
Base URL https://api.thetoken.dev/v1, your API key in TOKEN_API_KEY.
from openai import OpenAIimport os
client = OpenAI(base_url="https://api.thetoken.dev/v1", api_key=os.environ["TOKEN_API_KEY"])result = client.embeddings.create(model="baai/bge-m3", input="TOKEN serves embeddings")print(len(result.data[0].embedding))import OpenAI from "openai";
const client = new OpenAI({ baseURL: "https://api.thetoken.dev/v1", apiKey: process.env.TOKEN_API_KEY });const result = await client.embeddings.create({ model: "baai/bge-m3", input: "TOKEN serves embeddings" });console.log(result.data[0].embedding.length);curl https://api.thetoken.dev/v1/embeddings \ -H "Authorization: Bearer $TOKEN_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "baai/bge-m3", "input": "TOKEN serves embeddings"}'Providers
Requests go to the first provider and move down the list if it fails. TOKEN's price applies whichever provider serves the request.
| # | Provider | Precision (as the provider states it) | Data location | us-eu-hosted |
|---|---|---|---|---|
| 1 | DeepInfra | FP32 · checked October 19, 2026 | US | Yes |
| 2 | Novita AI | Not stated (backup only) · checked October 19, 2026 | US | Yes |
Availability
Request success: the share of real requests answered, after any failover; errors count against it. Last 30 days, one bar per day:
- 24 hours: …
- 7 days: …
- 30 days: …
Endpoint availability: a request to this model every 5 minutes from outside our infrastructure (Better Stack), on the status page. Method: Reliability.
Capabilities
| Input | text |
|---|---|
| Output | embedding |
| Reasoning | No |
| Tool calling | Not stated by the maker |
| Structured output (JSON schema) | Not stated by the maker |
| Maximum output | Not stated by the maker |
| Context window | 8,192 tokens |
As stated by BAAI, checked October 24, 2026: bge-model.com, github.com, huggingface.co.
Questions
Who makes BGE-M3?
BGE-M3 is made by BAAI and was released on January 30, 2024. TOKEN serves the maker's released weights through upstream providers; it does not train or modify the model.
What is the context window of BGE-M3?
8,192 tokens (8K).
At what precision does TOKEN serve BGE-M3?
The first provider in the routing order serves FP32, the maker's own native release format. TOKEN never routes to a quantized copy; a provider that does not state its precision is used only as a labelled backup.
Which providers serve BGE-M3 on TOKEN?
DeepInfra and Novita AI, in that routing order. If one fails, the request moves to the next.
How much does BGE-M3 cost?
$0.011 per 1M input tokens, the same whichever provider answers. You pay from prepaid credit in USD; each top-up carries a 5.5% credit purchase fee, shown at checkout.
Does BGE-M3 support tool calling, images or reasoning?
Tool calling: Not stated by the maker. Image input: No. Reasoning: No. Structured (JSON schema) output: Not stated by the maker.
Where are my requests processed?
At the provider that answers: US. Keys in the us-eu-hosted group only use US and EU hosts.
Does TOKEN store my prompts or responses?
No. TOKEN keeps token counts and request metadata for billing and logs, not the content of prompts or responses. See the privacy policy.
What happens if a provider of BGE-M3 fails?
TOKEN retries the request on the next provider in the routing order, up to three times, without any change on your side.
Will BGE-M3 be retired without notice?
No. TOKEN gives at least 30 days of notice before a model is retired.