# Gemini 4 Argon vs Step 5 Preview

> Side-by-side benchmark comparison for Gemini 4 Argon and Step 5 Preview across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/gemini-4-argon-vs-step-5-preview
- Last updated: September 30, 2026

- Shared sourced benchmarks: 4
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.7

## Quick Verdict

Pick Step 5 Preview if you want the stronger benchmark profile. Gemini 4 Argon only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- Step 5 Preview leads overall 68.55 to 64.59.
- The clearest category separation is in coding, where the averages are 68.4 for Gemini 4 Argon and 61.9 for Step 5 Preview.
- The biggest single benchmark swing is Terminal-Bench 4.0 in Agentic, with scores of 57.40% and 33.30%.
- Step 5 Preview is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- Step 5 Preview also has the larger context window at 1M.

## Model Snapshot

| Property | Gemini 4 Argon | Step 5 Preview |
|----------|----------|----------|
| Creator | Google | StepFun |
| Type | Proprietary | Pending |
| Reasoning | Reasoning | Reasoning |
| Context | null | 1M |
| Overall Score | 64.59 | 68.55 |
| Benchmarks Covered | 17 | 22 |
| Pricing (input/output) | $2.00 / $10.00 | $1.00 / $2.70 |

## Category Breakdown

### Agentic

- Winner: Directional only
- Gemini 4 Argon public-lane score: 63.7 (Estimated · #14/119)
- Step 5 Preview public-lane score: 63.3 (Estimated · #15/119)

| Benchmark | Gemini 4 Argon | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| AutomationBench | 51.3% | 44.0% | Gemini 4 Argon |
| Finance Agent v2 | 65.4% | Coming soon | Coming soon |
| Terminal-Bench 4.0 | 57.40% | 33.30% | Gemini 4 Argon |
| Agents' Last Exam | 39.5% | 29.5% | Gemini 4 Argon |
| OSWorld 2.0 | 69.2% | Coming soon | Coming soon |
| CWE-bench v1 | 68.0% | Coming soon | Coming soon |
| Terminal-Bench-Science 0.1 (6x verifier timeout) | 57.6% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 85.0% | Coming soon |
| CyberGym | Coming soon | 84.7% | Coming soon |
| Toolathlon-Verified | Coming soon | 74.1% | Coming soon |
| MCP Atlas | Coming soon | 85.6% | Coming soon |
| JobBench | Coming soon | 59.0% | Coming soon |
| APEX-Agents | Coming soon | 37.8% | Coming soon |
| DRACO | Coming soon | 83.3% | Coming soon |
| BrowseComp | Coming soon | 88.7% | Coming soon |

### Coding

- Winner: Directional only
- Gemini 4 Argon public-lane score: 68.4 (Supported · #8/144)
- Step 5 Preview public-lane score: 61.9 (Estimated · #16/144)

| Benchmark | Gemini 4 Argon | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| DeepSWE | 77.9% | 67.7% | Gemini 4 Argon |
| FrontierSWE v2 | 55.1% | Coming soon | Coming soon |
| Vibe Code Bench | 91.90% | Coming soon | Coming soon |
| PostTrainBench v1.1 | 45.3% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 85.0% | Coming soon |
| SciCode | Coming soon | 58.9% | Coming soon |
| ProgramBench | Coming soon | 80.5% | Coming soon |
| sweMarathon | Coming soon | 72.7% | Coming soon |
| MLS-Bench Lite | Coming soon | 40.5% | Coming soon |

### Reasoning

- Winner: Not comparable
- Gemini 4 Argon public-lane score: 77.1 (Unranked · 3 rankable rows)
- Step 5 Preview public-lane score: 79.6 (Unranked · 2 rankable rows)

| Benchmark | Gemini 4 Argon | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| Graphwalks BFS 128K | 99.7% | Coming soon | Coming soon |
| GraphWalks BFS 256K–1M | 84.2% | Coming soon | Coming soon |
| CritPt | Coming soon | 20.9% | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- Gemini 4 Argon public-lane score: Coming soon
- Step 5 Preview public-lane score: 61.2 (#28/49)

| Benchmark | Gemini 4 Argon | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| Chartography (no tools) | 71.6% | Coming soon | Coming soon |
| LVBench | 91.7% | Coming soon | Coming soon |
| OfficeQA Pro | Coming soon | 60.3% | Coming soon |
| MMMU-Pro | Coming soon | 76% | Coming soon |

### Knowledge

- Winner: Gemini 4 Argon
- Gemini 4 Argon public-lane score: 72.9 (Supported · #11/170)
- Step 5 Preview public-lane score: 67.5 (Supported · #19/170)

| Benchmark | Gemini 4 Argon | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| LABBench2 | 88.8% | Coming soon | Coming soon |
| GPQA-D | Coming soon | 93.5% | Coming soon |
| HLE | Coming soon | 46.5% | Coming soon |

### Instruction Following

- Winner: Not comparable
- Gemini 4 Argon public-lane score: Coming soon
- Step 5 Preview public-lane score: Coming soon

| Benchmark | Gemini 4 Argon | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| Gray Swan IPI (15 attempts) | 0.7% | Coming soon | Coming soon |

## FAQ

### Which is better overall, Gemini 4 Argon or Step 5 Preview?

Step 5 Preview is ahead overall on BenchLM right now.

### Where is the biggest gap between Gemini 4 Argon and Step 5 Preview?

The widest category gap is in coding, where the averages are 68.4 for Gemini 4 Argon and 61.9 for Step 5 Preview.

### How many benchmarks does Gemini 4 Argon cover on BenchLM?

Gemini 4 Argon currently has 17 sourced benchmark scores on BenchLM.

### How many benchmarks does Step 5 Preview cover on BenchLM?

Step 5 Preview currently has 22 sourced benchmark scores on BenchLM.

## Related Comparisons

- [Gemini 4 Argon vs GPT-6 Astra](/compare/gemini-4-argon-vs-gpt-6-astra)
- [Step 5 Preview vs GPT-6 Astra](/compare/gpt-6-astra-vs-step-5-preview)
- [Gemini 4 Argon vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-gemini-4-argon)
- [Step 5 Preview vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-step-5-preview)

## Explore More

- [Gemini 4 Argon profile](/models/gemini-4-argon)
- [Step 5 Preview profile](/models/step-5-preview)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)