bge-small-en-v1.5

Other · 1 offer

Free to try

Measured

From OpenVibe.Report: real calls in the last 30 days, with no prompt and no person in them.

Not measured yet: a model is measured after 20 calls in 30 days.

Every offer that serves it

Price in, out and cached per million tokens; speed and uptime from the last seven days of real requests; the privacy label and the provider’s own data policy.

Offer	Status	In / out per M	Speed	First token	Uptime	Errors	Privacy	Data policy	Try
OpenVibe Free to try	Up	free capacity	16 ms p50 · 105 ms p95	–	85.5%	14.5%	Private	not stated	Try it

Compare bge-small-en-v1.5 with another model →

Activity

Requests a day across every offer.

062026-11-06: 2 requests
072026-11-07: 5 requests
082026-11-08: 9 requests
092026-11-09: 2 requests
102026-11-10: 11 requests
112026-11-11: 59 requests
122026-11-12: 3 requests
132026-11-13: 3 requests

Measured performance

Measured runs on 2026-11-11
Provider	Task kind	Runs	Median latency (ms)	95th percentile (ms)	First token (ms)	Cost per run (USD)
local	embedding	59	101	278.2	–	0

Metrics and rating data

Bench rating: chat.arena

No single model rating has twenty qualifying games.

Source: OpenVibe.Bench. Latency and cost come from OpenVibe.AI receipts.

Presets

A preset fixes the router’s floor, ceiling and policy, and you name it as the model.

Cheapestmodel: "openvibe/economy"

The lowest cost per successful answer.

Fastestmodel: "openvibe/fast"

The quickest measured offer.

Privatemodel: "openvibe/private"

Only capacity OpenVibe runs or that sits on a local machine.

Balancedmodel: "openvibe/auto"

Cost, health and latency weighed together. The default.

These four aliases work today. A saved preset (model: "@preset/private") is the router’s floor, ceiling, quantization, region and data policy written down once. The private preset is the strict one: only offers that keep nothing and train on nothing, or OpenVibe’s own deployments, may answer, and the router refuses rather than fall back.

Try it

Chat with bge-small-en-v1.5 How it ranks

Use it from code

curl https://api.openvibe.exchange/v1/chat/completions \
  -H "Authorization: Bearer $OPENVIBE_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"local/bge-small-en-v1.5","messages":[{"role":"user","content":"Hello"}]}'

import OpenAI from 'openai';
const ai = new OpenAI({ baseURL: 'https://api.openvibe.exchange/v1', apiKey: process.env.OPENVIBE_KEY });
const r = await ai.chat.completions.create({ model: 'local/bge-small-en-v1.5', messages: [{ role: 'user', content: 'Hello' }] });

The endpoint, keys and usage: the API.