> Gemini 4 Argon vs Step 5 Preview: index scores ~188 and ~175 (GPT-4 = 100), API prices, context windows and results on the same independent tests, side by side.

Up to four models

# Compare

Side by side on the same independent tests, with price and context. The link in your address bar shares this exact comparison.

[Gemini 4 Argon](https://themodelindex.org/models/gemini-4-argon/)

[Step 5 Preview](https://themodelindex.org/models/step-5-preview/)

Add model

Next to Gemini 4 Argon

Cheapest at this level

Claude Sonnet 5.5

Best open model, measured

Kimi K3

Closest rival, another lab

GPT-6.1 Sol

Where it started

GPT-4

Presets

Best open vs best closed (measured)

Today vs GPT-4 (2023)

Top three labs

US vs China, best of each

[Gemini 4 Argon](https://themodelindex.org/models/gemini-4-argon/)

~188

estimated

Weights

Closed

Released

Sep 2026

Price in / out

–

Context

–

Core tests taken

–

[Step 5 Preview](https://themodelindex.org/models/step-5-preview/)

~175

estimated

Weights

Closed

Released

Sep 2026

Price in / out

–

Context

–

Core tests taken

–

Gemini 4 Argon and Step 5 Preview are too close to call (13 points apart, within the uncertainty) (one score is an estimate).

## Test by test · independent results · Epoch AI

Gemini 4 Argon

Step 5 Preview

These models share no independent tests yet.

Only tests that at least two of these models have taken. Bold rows are the core tests shown on every profile; all of these results inform the index score. Task length (METR) is listed below the chart because it is a duration, not a percentage.

---
Source: https://themodelindex.org/compare/gemini-4-argon-vs-step-5-preview/ · The Model Index · data as of 2026-10-01