Loading benchmark catalogue
Compare performance benchmarks, coding intelligence, reasoning evaluations, context limits, and API economics between Step 5 Preview (StepFun) and Gemini 4 Argon (Google DeepMind).
| Metric / Capability | Step 5 Preview | Gemini 4 Argon |
|---|---|---|
| Developer / Organization | StepFun | Google DeepMind |
| Release Date | 2026-10-08 | 2026-10-13 |
| Deployment / Weights | Open Weights | Proprietary Cloud API |
| Context Window | 1.0M tokens (1,048,576 tokens) | 1.0M tokens (1,048,576 tokens) |
| Input Pricing (/1M Tokens) | Free (Open Weights) | $2 |
| Output Pricing (/1M Tokens) | Free (Open Weights) | $10 |
| Benchmark / Domain | Step 5 Preview | Gemini 4 Argon | Advantage |
|---|---|---|---|
Artificial Analysis Intelligence Index Composite index aggregating 10+ evaluations of general reasoning and knowledge. | 43.7% | 52.6% | Gemini 4 Argon (+20%) |
Artificial Analysis Coding Agent Index v2.0 Composite index of coding-agent systems (model + harness) across FrontierCode v1.1, SWE-Atlas-QnA, and Terminal-Bench v4. | 43.1% | 53.2% | Gemini 4 Argon (+23%) |
Terminal-Bench 4.0 Sandboxed command-line tasks requiring shell and file-system operation. | 33.3% | 57.4% | Gemini 4 Argon (+72%) |
ARC-AGI-2 Novel visual grid puzzles requiring abstract rule induction. | — | 91.3% | — |
GPQA Diamond Graduate-level physics, chemistry, and biology questions written by domain experts. | 93.5% | — | — |
Humanity's Last Exam Expert-written questions across hundreds of disciplines designed to resist saturation. | 46.5% | — | — |
FrontierMath v2 (Tiers 1-3) Unpublished, research-level mathematics problems across Tiers 1-3. | — | 86.0% | — |
Design Arena ELO Human-preference ELO across UI, 3D, data-visualization, and game-dev generations. | — | 1438 | — |
Gemini 4 Argon outperforms Step 5 Preview across 3 evaluated benchmarks compared to 0 for Step 5 Preview, excelling in Artificial Analysis Intelligence Index, Artificial Analysis Coding Agent Index v2.0, Terminal-Bench 4.0.
Both models feature equivalent context windows of 1.0M tokens (1,048,576 tokens).
Step 5 Preview is more cost-effective. Input tokens cost $Free (Open Weights)/M tokens vs $2/M tokens for Gemini 4 Argon.
Step 5 Preview is an open-weights model. Gemini 4 Argon is a proprietary cloud API.
Analyze 130+ models side-by-side using interactive radar charts, cost-efficiency quadrant plots, and live dataset filters on the full NeuBench index.
Open Interactive Comparator