Vo
VoyageAI by MongoDB: rerank-2.5-lite
rerank-2.5-lite is a reranker optimized for both latency and quality, delivering a 7.16% improvement in retrieval accuracy over Cohere Rerank v3.5 across 93 datasets. It also outperformed Cohere Rerank v3.5...
- Input
- text
- Output
- rerank
Providers
Available through 1 free provider. Use the model ID and API settings below to connect.
Vercel AI Gateway
VAVercel AI Gateway
* The provider match may differ from this model. Check its docs for details.
Benchmarks
Compare scores and explore results by category.
No benchmark results are available for this model yet.
Frequently asked questions
Free access, setup, and what to expect.
Free API access for rerank-2.5-lite is listed through Vercel AI Gateway, but a specific request or token allowance isn’t available in the listed data. Check the rerank-2.5-lite provider pricing and rate limits for what’s included and any account requirements.
To use rerank-2.5-lite, choose one of the listed rerank-2.5-lite API providers and follow the provider’s account and API-key setup guide. Copy the rerank-2.5-lite model ID and connection settings for your chosen provider into your app or API client. Model IDs can differ between providers. API requests go directly to the provider; Free Models is a comparison site.
rerank-2.5-lite accepts text as listed input formats. Listed output formats include rerank. Image or audio input support does not imply image or audio generation. Check the chosen provider’s rerank-2.5-lite documentation for supported file types and request formats.
rerank-2.5-lite has a listed context window of 32K tokens. Tokens are small pieces of content rather than a word count. Instructions, conversation history, and other inputs use context space, and the response also needs room. The maximum output length for rerank-2.5-lite may be smaller than the context window and can vary by provider.
Neither reasoning nor tool calling support is listed for rerank-2.5-lite. Reasoning enables additional processing for complex problems. Tool calling allows a model to request functions that your app executes. Check the chosen provider’s documentation for capability availability and setup.
No benchmark results are listed for rerank-2.5-lite yet. Missing scores don’t establish poor performance. Compare scores within the same benchmark: coding indexes, Elo ratings, and win rates measure different things. The index comparison charts include only scored free models listed on Free Models. Evaluate rerank-2.5-lite on examples from your own task alongside context size and provider limits.