colqwen2.5 v0.2
An open model that searches document pages from their images. vidore/colqwen2.5-v0.2
- Accuracy
3 ViDoRe v3 sets, mean nDCG@10 · not measured
- Time per request
Not timed one request at a time
- Serverless price
Call it
from pathlib import Pathfrom urllib.request import urlopenfrom sie_sdk import SIEClientclient = SIEClient("http://localhost:8080")model = "modelvidore/colqwen2.5-v0.2"query = "queryDid JPMorganChase execute more than half of its planned $30 billion stock repurchase program by year-end?"pages = [ imagejpmorgan-2024-p106.jpgbrowse, imagejpmorgan-2024-p107.jpgbrowse, imagejpmorgan-2024-p108.jpgbrowse,]images = [urlopen(page).read() if page.startswith("https://") else Path(page).read_bytes() for page in pages]q = client.encode( model, [{"text": query}], output_types=["multivector"], is_query=True,)[0]["multivector"]docs = client.encode( model, [{"images": [image]} for image in images], output_types=["multivector"], is_query=False,)print(q)print(docs)import { readFile } from 'node:fs/promises';
import { SIEClient } from '@superlinked/sie-sdk';
const client = new SIEClient('http://localhost:8080');
const model = 'vidore/colqwen2.5-v0.2';
const query = "Did JPMorganChase execute more than half of its planned $30 billion stock repurchase program by year-end?";
const pages = [
"jpmorgan-2024-p106.jpg",
"jpmorgan-2024-p107.jpg",
"jpmorgan-2024-p108.jpg",
];
const sources = pages;
const images = await Promise.all(sources.map(async (source) => {
if (!source.startsWith('https://')) return readFile(source);
const response = await fetch(source);
if (!response.ok) throw new Error('Could not load the original page image.');
return new Uint8Array(await response.arrayBuffer());
}));
const [q] = await client.encode(
model, [{ text: query }],
{ outputTypes: ['multivector'], isQuery: true },
);
const docs = await client.encode(
model, images.map((image) => ({ images: [image] })),
{ outputTypes: ['multivector'], isQuery: false },
);
console.log(q, docs);page_0=$(base64 < 'jpmorgan-2024-p106.jpg' | tr -d '\n')
page_1=$(base64 < 'jpmorgan-2024-p107.jpg' | tr -d '\n')
page_2=$(base64 < 'jpmorgan-2024-p108.jpg' | tr -d '\n')
curl http://localhost:8080/v1/encode/vidore%2Fcolqwen2.5-v0.2 \
-H "Content-Type: application/json" \
-d "{\"items\":[{\"text\":\"Did JPMorganChase execute more than half of its planned \$30 billion stock repurchase program by year-end?\"}],\"params\":{\"output_types\":[\"multivector\"],\"options\":{\"is_query\":true}}}"
curl http://localhost:8080/v1/encode/vidore%2Fcolqwen2.5-v0.2 \
-H "Content-Type: application/json" \
-d "{\"items\":[{\"images\":[{\"data\":\"$page_0\",\"format\":\"jpeg\"}]},{\"images\":[{\"data\":\"$page_1\",\"format\":\"jpeg\"}]},{\"images\":[{\"data\":\"$page_2\",\"format\":\"jpeg\"}]}],\"params\":{\"output_types\":[\"multivector\"],\"options\":{\"is_query\":false}}}"Build the "Visual document search" capability into my app using the Superlinked Inference Engine (SIE).
Context
- SIE is an OpenAI-style inference API. Python SDK: `from sie_sdk import SIEClient`; TypeScript: `@superlinked/sie-sdk`.
- Base URL: http://localhost:8080, my own SIE server, which takes no API key.
- Model: vidore/colqwen2.5-v0.2 (SIE primitive: /encode). Keep the model id configurable.
Task
- Input: a text query plus one or more document page images.
- Behaviour: return native token vectors for the query and document page images
- Preserve the original page image bytes. Encode the text query with output_types ["multivector"] and is_query true, then all page images in one batch with output_types ["multivector"] and is_query false. Python: client.encode with output_types and is_query. TypeScript: client.encode with outputTypes and isQuery. Return the native token vectors without application scoring or ranking.
Deliverables
- A typed client wrapper, an application-level function for this task, error handling for timeouts/empty input, and unit tests with a stubbed client.
- Wire it into my existing stack (ask me which framework if unclear) and add a short usage example.Next to the other page search models
| Model | 3 ViDoRe v3 sets, mean nDCG@10 | Time per request | Serverless price | Size |
|---|---|---|---|---|
| colqwen2.5 v0.2 vidore | Not measured | Not timed | Run in your cloud | 7.0B |
| Tomoro ColQwen3 4B TomoroAI | Not measured | Not timed | $2.00 per 1M tokens $0.20 per 1,000 images | 4.4B |
| nemotron colembed vl 4b v2 nvidia | 0.720 | Not timed | Run in your cloud | 4.8B |
| v splade quality naver | 0.527 | Not timed | Run in your cloud | 330M |
| colpali v1.3 hf vidore | Not measured | Not timed | Run in your cloud | 2.9B |
Facts
- Size
- 7.0B parameters
- License
- MIT
- Reads
- Text (up to 2,048 tokens), image
- Endpoint
-
encode
- Accuracy
- Run on SIE against the public dataset named in each figure
- 3 ViDoRe v3 sets
- The mean over Vidore3ComputerScienceRetrieval, Vidore3FinanceEnRetrieval and Vidore3PharmaceuticalsRetrieval
- Prices
- Serverless list prices, effective 10 Nov 2026
Run colqwen2.5 v0.2 in your own cloud, or start on Serverless
SIE Serverless
Task-specific models for your agents, with no GPUs to operate.
- No idle GPUs, pay for what you use
- Fits your stack: SDK, API, CLI, MCP
- Self-host the same open-source engine
- SOC 2 Type 2, US or EU data residency
no credit card required
apply for $500 more
Native Inference
Kubernetes in your own cloud, backed by our engineers.
- Terraform modules for your cloud
- Apache-2.0, same engine as Serverless
- GPU workers scale to zero between jobs
- Separate worker pools for each tenant
Deploy SIE to our AWS account with the superlinked/sie/aws Terraform module. Docs: superlinked.com/docs/deployment.Deploy SIE to our Google Cloud project with the superlinked/sie/google Terraform module. Docs: superlinked.com/docs/deployment.Deploy SIE to our Azure subscription with the superlinked/sie/azure Terraform module. Docs: superlinked.com/docs/deployment.Deploy SIE to our Alibaba Cloud account with the superlinked/sie/alicloud Terraform module. Docs: superlinked.com/docs/deployment. SIE Local
The open-source server on your own machine.
- Runs on your NVIDIA GPU
- One container, no cluster
- Process your data fully offline
- Same SDK and model IDs as Serverless
docker run --gpus all -p 8080:8080 ghcr.io/superlinked/sie-server:latest-cuda12-default