Skip to main content
Watch: I Built an AI App Without Leaving Databricks

Qwen rerankers - verified

Qwen reranker models

Qwen (open weights): 3 reranker models, read from the provider's own pages on October 18, 2026, 3 with open weights. Qwen rerankers are self-hosted from open weights, with no per-call price.

Read from huggingface.co. Compare every reranker

Models
3
Current
3
Older
0
Open weights
3

Billing and caveats

Qwen (open weights)

32k is the card's rounded figure; it is recorded here as 32000. Alibaba's hosted qwen3-rerank API is listed separately under Alibaba Cloud Model Studio.

Models

Qwen3-Reranker-0.6B

Qwen/Qwen3-Reranker-0.6B - Open weights - Self-host (open weights or NIM containers)

GAinferred
Price (provider's unit)
Not stated

No per-call price: self-hosted.

Status

GAinferred

Status inferred: the source lists the model but does not label it.

Context

32,000 tokens

context length (card says 32k)

“- Context Length: 32k”
huggingface.co
Documents per request
Not stated
Languages

100+ languages

“- Supported Languages: 100+ Languages”
huggingface.co
Open weights

Yes - Apache 2.0, 596M parameters

Qwen/Qwen3-Reranker-0.6B - license

“"license":"apache-2.0"”
huggingface.co
“"total":595776512”
595,776,512 parameters - huggingface.co

Instruction-aware; the card recommends task-specific instructions written in English.

Qwen3-Reranker-4B

Qwen/Qwen3-Reranker-4B - Open weights - Self-host (open weights or NIM containers)

GAinferred
Price (provider's unit)
Not stated

No per-call price: self-hosted.

Status

GAinferred

Status inferred: the source lists the model but does not label it.

Context

32,000 tokens

context length (card says 32k)

“- Context Length: 32k”
huggingface.co
Documents per request
Not stated
Languages

100+ languages

“- Supported Languages: 100+ Languages”
huggingface.co
Open weights

Yes - Apache 2.0, 4.02B parameters

Qwen/Qwen3-Reranker-4B - license

“"license":"apache-2.0"”
huggingface.co
“"total":4021784576”
4,021,784,576 parameters - huggingface.co

Instruction-aware; 36 layers per the series table.

Qwen3-Reranker-8B

Qwen/Qwen3-Reranker-8B - Open weights - Self-host (open weights or NIM containers)

GAinferred
Price (provider's unit)
Not stated

No per-call price: self-hosted.

Status

GAinferred

Status inferred: the source lists the model but does not label it.

Context

32,000 tokens

context length (card says 32k)

“- Context Length: 32k”
huggingface.co
Documents per request
Not stated
Languages

100+ languages

“- Supported Languages: 100+ Languages”
huggingface.co
Open weights

Yes - Apache 2.0, 8.19B parameters

Qwen/Qwen3-Reranker-8B - license

“"license":"apache-2.0"”
huggingface.co
“"total":8188548096”
8,188,548,096 parameters - huggingface.co

Instruction-aware; largest of the three sizes.

FAQ

How much does Qwen reranking cost?

Qwen3-Reranker-0.6B: no per-call price (self-hosted). Qwen3-Reranker-4B: no per-call price (self-hosted). Qwen3-Reranker-8B: no per-call price (self-hosted). Prices are in Qwen's own units as of October 18, 2026.

What is the context length of Qwen rerankers?

Qwen3-Reranker-0.6B: 32,000 tokens, context length (card says 32k). Qwen3-Reranker-4B: 32,000 tokens, context length (card says 32k). Qwen3-Reranker-8B: 32,000 tokens, context length (card says 32k).

Can I run Qwen rerankers myself?

Yes for Qwen3-Reranker-0.6B (Apache 2.0), Qwen3-Reranker-4B (Apache 2.0) and Qwen3-Reranker-8B (Apache 2.0): the weights are on Hugging Face.

Sources

Last verified from each provider's own docs, pricing pages and Hugging Face cards; every quote was string-matched against the page that day. Prices are shown in the provider's own unit and change without notice, so check the linked source before you commit a budget.

Newsletter

Know when a reranker API changes

One email per week. New model releases, retrieval and pricing updates and what they mean for what you build.