NV Embed V2
NVIDIA's NV Embed V 2 is a generalist embedding model that excels at text embedding tasks, as evidenced by its top ranking on the Massive Text Embedding Benchmark with a score of 72.31. It features a base decoder-only LLM architecture, specifically Mistral-7B-v0.1, and utilizes a latent-attention pooling type to produce high-quality embeddings. Notably, NV Embed V 2 incorporates a novel two-staged instruction tuning method and hard-negative mining approach to enhance its accuracy, particularly in retrieval tasks.
Input
Output
Context
33K
Max Output
-
Parameters
-
Technical Specifications
Input Modalities
Output Modalities
Capabilities
Loading capabilities…
Resources & Links
More models by NVIDIA
The creator's other models in the catalog, with their context, size and license where published.
Compare it, or keep browsing
Put this model beside its alternatives on the same evidence, or go back to the full catalog.
Frequently asked questions about NV Embed V2
Answered from the entry's own fields: context, license, modalities, evidence, serving and the memory to self-host.
NV Embed V2 advertises a context window of 32,768 tokens. The figure is the creator's published maximum; a given host may serve less, and the gateway routes on what each host actually serves.
Yes. NV Embed V2 is an open-weight model released under the Cc by nc 4.0 license, so the weights can be downloaded and self-hosted within that license's terms.
NV Embed V2 accepts Text and produces Embedding. The capabilities card on this page lists which API features each deployment honours, such as function calling and structured output, with the source each was checked against.
No published benchmark result for NV Embed V2 is in the catalog yet, so the model is shown as unmeasured. It is not ranked or estimated; the router treats it as unknown for every task until a suite measures it.
Not on the managed pool today; NV Embed V2 is listed for reference and comparison. Connect your own provider key or endpoint that serves it and the gateway runs it on your account, with routing decisions recorded the same way.
NVIDIA has not published a parameter count for NV Embed V2, so the catalog cannot estimate its memory footprint. The capacity planner can size it from a parameter count you supply.
Fields collected from public registries, host APIs and benchmark publishers, each tagged with its source.
Last updated: Aug 28, 2026
Put intelligence in the middle.
One gateway in front of every model, with your policies applied and every decision on record. Start with $5 of credit and 5,000 routing decisions a month, no card required.