We use cookies. Cookies and browser storage keep you signed in and remember your preferences. With your permission we also measure how the site is used. You can any time. See our cookie policy.

New here? See how it works A 1-minute tour of the playground, models and your API key.

Hermes 4 405B vs Relace Search: API price, context & features · api14

Nous: Hermes 4 405B vs Relace: Relace Search

A side-by-side comparison of Nous: Hermes 4 405B ($3.50 / 1M tokens) and Relace: Relace Search ($3.50 / 1M tokens) using current USD prices and capabilities from the api14 catalog. Both run through the same OpenAI-compatible API.

Specifications

	Nous: Hermes 4 405B	Relace: Relace Search
Provider	Nous	Relace
Input	$3.50 / 1M tokens	$3.50 / 1M tokens
Output	$10.50 / 1M tokens	$10.50 / 1M tokens
Cache read	$0.70 / 1M tokens	-
Context window	131,072	256,000
Max output	117,964	128,000
Input	text	text
Output	text	text
Tool calling
Structured output
Reasoning control
Added	Aug 26, 2025	Dec 8, 2025

Key differences

Both have about the same input price.
Relace: Relace Search reads 2.0× more context (256,000 tokens).
Only Relace: Relace Search supports tool calling.
Only Nous: Hermes 4 405B supports structured JSON output.
Only Nous: Hermes 4 405B supports reasoning effort control.
A request with 1,000 input and 500 output tokens costs $0.00875 with Nous: Hermes 4 405B and $0.00875 with Relace: Relace Search.

In depth

Nous: Hermes 4 405B is a text model from Nous, released on 2025-08-26. Relace: Relace Search is a text model from Relace, released on 2025-12-08. Both are served through api14 as text in, text out models, so neither adds image, audio or video handling to your pipeline. The choice between them comes down to context size, tool use, structured output and reasoning rather than to any difference in modality.

The clearest split is context length. Relace Search supports 256000 tokens, roughly double the 131072 tokens of Hermes 4 405B, so it is the one to reach for when you need to keep very long documents, large codebases or extended conversation histories in a single request. Relace Search is also the only one of the two that accepts tools and tool_choice, which makes it the natural fit for agent style workflows where the model has to call functions or external services. Hermes 4 405B goes the other way: it is the only one that exposes structured outputs and response_format alongside reasoning and include_reasoning, plus a wider sampling surface that includes frequency_penalty, presence_penalty, repetition_penalty, logprobs, top_logprobs, top_k, temperature and top_p. If you need the model to emit a schema conforming object, or you want to inspect token level probabilities, or you want visible reasoning traces, Hermes 4 405B is the model that supports it. On price the two are about the same, so cost is not the deciding factor here.

For day to day work, pick Relace Search when the task is retrieval heavy or tool driven: searching across a large corpus, feeding in long transcripts, or letting the model decide when to call a function. Its larger context and tool support line up with those jobs. Pick Hermes 4 405B when the task is generation under a strict contract: extracting fields into JSON, classifying with a fixed label set, or any prompt where you want reasoning output and fine grained control over sampling. The extra parameters also help when you are tuning for diversity or repetition, since Relace Search only exposes max_tokens, response_format, seed, stop, temperature, tool_choice, tools and top_p.

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 message structure, and the only thing that changes is the model id. Use "nousresearch/hermes-4-405b" for Hermes 4 405B and "relace/relace-search" for Relace Search. Everything else in your client, including the chat completions shape and the way you read the response, stays identical, so switching between them is a one line change and you can route different requests to different models from the same code path.

There are cases where the other model is still the better pick. If you are already using Relace Search for a long context tool calling task but you need the final answer formatted as structured output, Hermes 4 405B is the one that supports structured outputs, so a two step flow where Relace Search gathers and Hermes 4 405B formats can make sense. Conversely, if you are using Hermes 4 405B and your prompt grows past its 131072 token window, Relace Search is the model that can still hold it.

Switch with one string

Both models use the same endpoint and request shape. Change the model value from "nousresearch/hermes-4-405b" to "relace/relace-search" and prompts, streaming and tools keep working.

Nous: Hermes 4 405B details & playground →Relace: Relace Search details & playground →

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api14.com/api/v1",
  apiKey: process.env.API14_API_KEY,
});

const result = await client.chat.completions.create({
  "model": "relace/relace-search",
  "messages": [
    {
      "role": "user",
      "content": "Explain vector databases in two sentences."
    }
  ]
});
console.log(result.choices[0].message.content);

Nous: Hermes 4 405B vs Relace: Relace Search

Specifications

Key differences

In depth

Switch with one string

More comparisons