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

NVIDIA NIM rerankers - verified

NVIDIA NIM reranker models

NVIDIA NIM: 3 reranker models, read from the provider's own pages on October 18, 2026, 2 with open weights. NVIDIA NIM rerankers are self-hosted from open weights, with no per-call price.

Read from docs.nvidia.com, huggingface.co. Compare every reranker

Models
3
Current
3
Older
0
Open weights
2

Billing and caveats

NVIDIA NIM

NIM containers are self-hosted; the support matrix lists no per-call price, so pricing is null.

Models

Llama Nemotron Rerank 1B v2

nvidia/llama-nemotron-rerank-1b-v2 - Self-hosted NIM and open weights - Self-host (open weights or NIM containers)

GA
Price (provider's unit)
Not stated

No per-call price: self-hosted.

Status

GA

“This model is ready for commercial use.”
huggingface.co
Context

8,192 tokens

query plus passage, optimized NIM profiles

“nvidia/llama-nemotron-rerank-1b-v2 · [Model card](https://build.nvidia.com/nvidia/llama-nemotron-rerank-1b-v2/modelcard) · 8192”
docs.nvidia.com
Documents per request
Not stated
Languages

Evaluated on 26 languages

“This model was evaluated on 26 languages”
huggingface.co
Open weights

Yes - OpenMDW-1.1, 1.24B parameters

nvidia/llama-nemotron-rerank-1b-v2 - license

“Use of this model is governed by the [OpenMDW License Agreement, version 1.1 (OpenMDW-1.1)](https://openmdw.ai/license/1-1/).”
huggingface.co
“"total":1235816448”
1,235,816,448 parameters - huggingface.co

Non-optimized NIM profile drops the limit to 4096 tokens. Card adds the Llama 3.2 Community License as additional terms.

Llama Nemotron Rerank VL 1B v2

nvidia/llama-nemotron-rerank-vl-1b-v2 - Self-hosted NIM and open weights - Self-host (open weights or NIM containers)

GA
Price (provider's unit)
Not stated

No per-call price: self-hosted.

Status

GA

“This model is ready for commercial use.”
huggingface.co
Context

8,192 tokens

optimized NIM profiles

“nvidia/llama-nemotron-rerank-vl-1b-v2 · [Model card](https://build.nvidia.com/nvidia/llama-nemotron-rerank-vl-1b-v2/modelcard) · 8192”
docs.nvidia.com
Documents per request
Not stated
Languages

Multilingual (no count stated)

“"language":["multilingual"]”
huggingface.co
Open weights

Yes - NVIDIA Open Model License, 1.68B parameters

nvidia/llama-nemotron-rerank-vl-1b-v2 - license

“The use of this model is governed by the [NVIDIA Open Model License Agreement](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/)”
huggingface.co
“"total":1678256576”
1,678,256,576 parameters - huggingface.co

Multimodal (image, text or both) document reranker, released 12/18/2025. The Hugging Face card says "The model's max sequence length is 10240" while the NIM support matrix lists 8192.

Llama Nemotron Rerank 500m v2

nvidia/llama-nemotron-rerank-500m-v2 - Self-hosted NIM container - Self-host (open weights or NIM containers)

GAinferred
Price (provider's unit)
Not stated

No per-call price: self-hosted.

Status

GAinferred

“Reranking NIM supports the following models.”
Status inferred from - docs.nvidia.com
Context

8,192 tokens

optimized NIM profiles

“nvidia/llama-nemotron-rerank-500m-v2 · [Model card](https://build.nvidia.com/nvidia/llama-3_2-nemoretriever-500m-rerank-v2/modelcard) · 8192”
docs.nvidia.com
Documents per request
Not stated
Languages
Not stated
Open weights
Not stated

Note: No Hugging Face repo was found at nvidia/llama-3.2-nemoretriever-500m-rerank-v2 (the API returned 401); open-weight status unverified.

Non-optimized NIM profile drops the limit to 4096 tokens.

FAQ

How much does NVIDIA NIM reranking cost?

Llama Nemotron Rerank 1B v2: no per-call price (self-hosted). Llama Nemotron Rerank VL 1B v2: no per-call price (self-hosted). Llama Nemotron Rerank 500m v2: no per-call price (self-hosted). Prices are in NVIDIA NIM's own units as of October 18, 2026.

What is the context length of NVIDIA NIM rerankers?

Llama Nemotron Rerank 1B v2: 8,192 tokens, query plus passage, optimized NIM profiles. Llama Nemotron Rerank VL 1B v2: 8,192 tokens, optimized NIM profiles. Llama Nemotron Rerank 500m v2: 8,192 tokens, optimized NIM profiles.

Can I run NVIDIA NIM rerankers myself?

Yes for Llama Nemotron Rerank 1B v2 (OpenMDW-1.1) and Llama Nemotron Rerank VL 1B v2 (NVIDIA Open Model License): the weights are on Hugging Face. The others are not published as open weights in the sources checked.

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.