Rerank v3.5
Reranker
Cohere's flagship reranker. Re-orders retrieved passages by relevance, the standard step-2 of high-quality RAG.
Rerank v3.5 is a embedding AI model from Cohere. It costs $2.000 per million input tokens.
INPUT
$2.000/M
per million input tokens
OUTPUT
—
per million output tokens
CONTEXT
—
no context window
Benchmarks
No published benchmark scores tracked for this model yet. Frontier and reasoning models from major providers have scores; smaller models and inference-host variants typically inherit the underlying open-weights score.
More from Cohere
See all 12 →Command A+
Frontier · 256,000 ctx
in $2.500/Mout $10.000/M
Command A
Previous · 256,000 ctx
in $2.500/Mout $10.000/M
Command R+
Previous · 128,000 ctx
in $2.500/Mout $10.000/M
Command R
Balanced · 128,000 ctx
in $0.150/Mout $0.600/M
Command R7B
Compact · 128,000 ctx
in $0.037/Mout $0.150/M
Aya Expanse 32B
Multilingual · 128,000 ctx
in $0.500/Mout $1.500/M
Frequently asked questions
How much does Rerank v3.5 cost?
Rerank v3.5 costs $2.000 per million input tokens.
What is Rerank v3.5 best for?
Cohere's flagship reranker. Re-orders retrieved passages by relevance, the standard step-2 of high-quality RAG.
Who makes Rerank v3.5?
Rerank v3.5 is developed and served by Cohere.