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.

Command A vs Gemini 4 Argon: API price, context & features · api14

Cohere: Command A vs Google: Gemini 4 Argon

A side-by-side comparison of Cohere: Command A ($8.75 / 1M tokens) and Google: Gemini 4 Argon ($7.00 / 1M tokens) using current USD prices and capabilities from the api14 catalog. Both run through the same OpenAI-compatible API.

Specifications

	Cohere: Command A	Google: Gemini 4 Argon
Provider	Cohere	Google
Input	$8.75 / 1M tokens	$7.00 / 1M tokens
Output	$35.00 / 1M tokens	$35.00 / 1M tokens
Cache read	-	$0.35 / 1M tokens
Cache write	-	$0.66 / 1M tokens
Reasoning	-	$35.00 / 1M tokens
Audio input	-	$7.00 / 1M tokens
Image	-	$0.000007 / image
Web search	-	$49.00 / 1K searches
Context window	256,000	1,048,576
Max output	8,192	943,718
Input	text	audio, file, image, text, video
Output	text	text
Tool calling
Structured output
Reasoning control
Added	Mar 13, 2025	Oct 13, 2026

Key differences

Google: Gemini 4 Argon is 20% cheaper on input tokens.
Google: Gemini 4 Argon reads 4.1× more context (1,048,576 tokens).
Only Google: Gemini 4 Argon supports reasoning effort control.
A request with 1,000 input and 500 output tokens costs $0.03 with Cohere: Command A and $0.02 with Google: Gemini 4 Argon.

In depth

Cohere: Command A is a text model from Cohere, released on 13 March 2025. Google: Gemini 4 Argon is a text model from Google, released on 13 October 2026. Both are served through api14 under the model ids cohere/command-a and google/gemini-4-argon, and both are listed as text category models, so at the surface level they occupy the same slot in a model picker.

The differences show up in context, modalities and parameters. Gemini 4 Argon has the larger context window, roughly a million tokens against Command A's 256,000, and it accepts audio, file, image, text and video as input while Command A accepts text only. Both return text. Gemini 4 Argon is the only one of the two that exposes tool calling and reasoning controls (include_reasoning, reasoning, reasoning_effort, tool_choice, tools), while Command A carries sampling controls such as frequency_penalty, presence_penalty, top_k and seed. Structured outputs appear on both, so that is not a deciding factor. On price, Gemini 4 Argon is the cheaper option for input, and the two are about the same for output.

For a developer choosing between them, the practical split follows those facts. If your task is plain text in and text out, with a prompt that fits comfortably inside 256,000 tokens, Command A is a reasonable default, and its penalty and top_k parameters give you finer control over sampling behaviour. If you need to send images, audio, video or files alongside the prompt, or you need a context window well beyond a quarter of a million tokens, Gemini 4 Argon is the one that can accept that input at all. If your application relies on tool calling or on exposing reasoning effort as a knob, Gemini 4 Argon is again the only one of the two that supports it. Cost matters too: for input-heavy workloads Gemini 4 Argon is cheaper, while output cost is about the same on both, so a generation-heavy, text-only workload does not gain much from switching on price alone.

Calling either model through api14 works the same way, because the API is OpenAI-compatible. You point your existing client at the api14 base URL, send a chat completions request with your messages, and set the model field to either cohere/command-a or google/gemini-4-argon. Switching between the two is a one-line change to that model id; the request and response shapes stay the same. The parameters you pass should match what each model supports, so tool and reasoning fields belong on requests to google/gemini-4-argon, and penalty or top_k fields belong on requests to cohere/command-a.

There are still cases where Command A is the better pick even though Gemini 4 Argon is larger and cheaper on input. If your workload is text only and you want the sampling controls that Command A exposes, or you simply prefer Cohere as a provider, Command A covers that ground without the extra modality surface. Neither model is a strict superset of the other in the parameter list, so the choice comes down to whether you need multimodal input, a very large context or reasoning and tool controls, in which case pick Gemini 4 Argon, or a text-only model with penalty and top_k tuning, in which case Command A is enough.

Switch with one string

Both models use the same endpoint and request shape. Change the model value from "cohere/command-a" to "google/gemini-4-argon" and prompts, streaming and tools keep working.

Cohere: Command A details & playground →Google: Gemini 4 Argon 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": "google/gemini-4-argon",
  "messages": [
    {
      "role": "user",
      "content": "Explain vector databases in two sentences."
    }
  ]
});
console.log(result.choices[0].message.content);

Cohere: Command A vs Google: Gemini 4 Argon

Specifications

Key differences

In depth

Switch with one string

More comparisons