Cohere: Rerank 4 Pro vs Cohere: Rerank v3.5
A side-by-side comparison of Cohere: Rerank 4 Pro (Free) and Cohere: Rerank v3.5 (Free) using current USD prices and capabilities from the api14 catalog. Both run through the same OpenAI-compatible API.
Specifications
| Cohere: Rerank 4 Pro | Cohere: Rerank v3.5 | |
|---|---|---|
| Provider | Cohere | Cohere |
| Context window | 32,768 | 4,096 |
| Max output | 29,491 | 3,686 |
| Input | text | text |
| Output | rerank | rerank |
| Tool calling | ||
| Structured output | ||
| Reasoning control | ||
| Added | Apr 6, 2026 | Apr 5, 2026 |
Key differences
- Cohere: Rerank 4 Pro reads 8.0× more context (32,768 tokens).
In depth
Cohere: Rerank 4 Pro and Cohere: Rerank v3.5 are both reranking models made by Cohere. Reranking models take a query and a set of candidate documents or passages and return a relevance ordering, which makes them useful as a second stage in a retrieval pipeline after an initial vector or keyword search. Both models are listed on api14 under the same provider and the same category, and both accept text as input and produce a rerank result as output. Neither model exposes any supported parameters in the catalogue, so the main decision between them comes down to context capacity rather than tuning options.
The clearest difference between the two is context length. Rerank 4 Pro accepts up to 32768 tokens, while Rerank v3.5 accepts up to 4096 tokens. That is a substantial gap, and it is the only qualitative difference recorded between them: the bigger context belongs to Rerank 4 Pro. Both are text in and rerank out, so there is no difference in modalities to weigh. Neither model is flagged as the only one supporting tools, structured output or reasoning, so those axes do not separate them. The catalogue does not record a meaningful price difference in either direction, so cost is not a deciding factor here; check the live price table for current numbers.
In practice, the choice follows from how much text each candidate item carries. If you are reranking short passages, single paragraphs, or search snippets that comfortably fit inside a few thousand tokens, Rerank v3.5 is a reasonable fit and its smaller window is not a limitation. If your candidates are long documents, multi-section pages, or chunks that you have deliberately kept large to preserve surrounding context, Rerank 4 Pro is the better pick because it can take those inputs without truncation. The same logic applies when you want to pass a query plus several candidate documents in one request: the larger window of Rerank 4 Pro gives you more room before you have to split or trim anything.
Calling either model through api14 works the same way because the API is OpenAI-compatible. You send your request to the same endpoint with the same request shape, and the only thing that changes is the model id. Use "cohere/rerank-4-pro" when you want the larger context window, and "cohere/rerank-v3.5" when the smaller window is enough. Because both models list no supported parameters, you do not need to adjust any model-specific options when switching; the query and documents you pass are what drive the result. This makes it straightforward to start on one model and move to the other if your input sizes change.
There is still a case for Rerank v3.5 even when you have access to Rerank 4 Pro. If your pipeline is built around short chunks and you have no need for the larger window, the smaller model keeps your inputs within a tighter bound and matches the shape of the data you are already sending. Since the catalogue records no meaningful price difference and no capability difference beyond context length, the decision is essentially about input size, and Rerank v3.5 remains a valid choice for short-text reranking workloads.
Switch with one string
Both models use the same endpoint and request shape. Change the model value from "cohere/rerank-4-pro" to "cohere/rerank-v3.5" and prompts, streaming and tools keep working.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api14.com/api/v1",
apiKey: process.env.API14_API_KEY,
});
const result = await client.post("/rerank", { body: {
"model": "cohere/rerank-v3.5",
"query": "What is the capital of France?",
"documents": [
"Paris is the capital of France.",
"Berlin is in Germany."
],
"top_n": 1
} });
console.log(result.results);
More comparisons
- Cohere: Rerank 5 Fast vs Cohere: Rerank 5 Pro
- Cohere: Rerank 5 Fast vs VoyageAI by MongoDB: rerank-3
- Cohere: Rerank 5 Fast vs VoyageAI by MongoDB: rerank-3-lite
- Cohere: Rerank 5 Fast vs Qwen3 Reranker 8B
- Cohere: Rerank 5 Pro vs VoyageAI by MongoDB: rerank-3
- Cohere: Rerank 5 Pro vs VoyageAI by MongoDB: rerank-3-lite
- Cohere: Rerank 5 Pro vs Qwen3 Reranker 8B
- VoyageAI by MongoDB: rerank-3 vs VoyageAI by MongoDB: rerank-3-lite