Together AI
Rerank is dedicated-endpoint only on 2026-10-18; there is no per-token rerank price.
Together AI rerankers - verified
Together AI reranker models
Together AI: 2 reranker models, read from the provider's own pages on October 18, 2026, 1 with open weights. Together AI rerankers are run only on dedicated endpoints, $5.49 per GPU per hour (dedicated endpoint) (NVIDIA HGX H100 on-demand).
Read from docs.together.ai, together.ai, huggingface.co. Compare every reranker
- Models
- 2
- Current
- 2
- Older
- 0
- Open weights
- 1
Billing and caveats
Models
mxbai-rerank-large-v2 (Together dedicated)
mixedbread-ai/mxbai-rerank-large-v2 - API - Dedicated endpoint only (Together AI)
- Price (provider's unit)
- Not stated
“Rerank models like mxbai-rerank-large-v2 are only available for [dedicated model inference](https://api.together.ai/endpoints/configure).”
docs.together.ai Note: No serverless rerank price: rerank runs on dedicated endpoints billed per GPU hour, and the pricing page's Rerank tab rendered no models.
Dedicated capacity
$5.49 per GPU per hour (dedicated endpoint), NVIDIA HGX H100 on-demand
“NVIDIA HGX H100 · $5.49”
together.ai - Status
GAinferred
“Rerank models like mxbai-rerank-large-v2 are only available for dedicated model inference.”
Status inferred from - docs.together.ai - Context
8,000 tokens
per document (API-level statement)
“Long 8K context per document.”
docs.together.ai - Documents per request
- Not stated
- Languages
- Not stated
- Open weights
Yes - Apache 2.0, 1.54B parameters
mixedbread-ai/mxbai-rerank-large-v2 - license
“"license":"apache-2.0"”
huggingface.co “"total":1543714304”
1,543,714,304 parameters - huggingface.co
Same weights as the Mixedbread open model; Together hosts it only on dedicated endpoints.
- Together rerank docs: docs.together.ai/docs/rerank-overview
- Together pricing: together.ai/pricing
Llama-Rank-V1 (Together dedicated)
Salesforce/Llama-Rank-V1 - API - Dedicated endpoint only (Together AI)
- Price (provider's unit)
- Not stated
“Rerank models like mxbai-rerank-large-v2 are only available for [dedicated model inference](https://api.together.ai/endpoints/configure).”
docs.together.ai Note: No serverless rerank price: rerank runs on dedicated endpoints billed per GPU hour, and the pricing page's Rerank tab rendered no models.
Dedicated capacity
$5.49 per GPU per hour (dedicated endpoint), NVIDIA HGX H100 on-demand
“NVIDIA HGX H100 · $5.49”
together.ai - Status
GAinferred
“The following JSON data format with rank_fields is only supported on [dedicated model inference](/docs/dedicated-endpoints/overview) running the Salesforce/Llama-Rank-V1 model.”
Status inferred from - docs.together.ai - Context
8,000 tokens
per document (API-level statement)
“Long 8K context per document.”
docs.together.ai - Documents per request
- Not stated
- Languages
- Not stated
- Open weights
- Not stated
Note: Together's docs do not state whether Salesforce publishes these weights.
Only model on Together that accepts JSON documents with rank_fields.
- Together rerank docs: docs.together.ai/docs/rerank-overview
- Together pricing: together.ai/pricing
FAQ
How much does Together AI reranking cost?
mxbai-rerank-large-v2 (Together dedicated): no per-call price; dedicated NVIDIA HGX H100 on-demand $5.49 per GPU per hour (dedicated endpoint). Llama-Rank-V1 (Together dedicated): no per-call price; dedicated NVIDIA HGX H100 on-demand $5.49 per GPU per hour (dedicated endpoint). Prices are in Together AI's own units as of October 18, 2026.
What is the context length of Together AI rerankers?
mxbai-rerank-large-v2 (Together dedicated): 8,000 tokens, per document (API-level statement). Llama-Rank-V1 (Together dedicated): 8,000 tokens, per document (API-level statement).
Can I run Together AI rerankers myself?
Yes for mxbai-rerank-large-v2 (Together dedicated) (Apache 2.0): 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.