# Claude Opus 4.8 vs Step 5 Preview

> Side-by-side benchmark comparison for Claude Opus 4.8 and Step 5 Preview across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/claude-opus-4-8-vs-step-5-preview
- Last updated: September 21, 2026

- Shared sourced benchmarks: 5
- HTML indexing: indexable
- Ranking lane: BenchAlign v5

## Quick Verdict

Pick Claude Opus 4.8 if you want the stronger benchmark profile. Step 5 Preview only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- Claude Opus 4.8 leads overall 72.17 to 72.
- The clearest category separation is in multimodal & grounded, where the averages are 87.7 for Claude Opus 4.8 and 61.2 for Step 5 Preview.
- The biggest single benchmark swing is HLE in Knowledge, with scores of 57.9% and 46.5%.
- Step 5 Preview is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.

## Model Snapshot

| Property | Claude Opus 4.8 | Step 5 Preview |
|----------|----------|----------|
| Creator | Anthropic | StepFun |
| Type | Proprietary | Pending |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 1M |
| Overall Score | 72.17 | null |
| Benchmarks Covered | 38 | 22 |
| Pricing (input/output) | $5.00 / $25.00 | $1.00 / $2.70 |

## Category Breakdown

### Agentic

- Winner: Coming soon
- Claude Opus 4.8 public-lane score: 63.3 (Supported · #13/154)
- Step 5 Preview public-lane score: Coming soon

| Benchmark | Claude Opus 4.8 | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 3.0 | 21.1% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 74.6% | Coming soon | Coming soon |
| BrowseComp | 84.3% | 88.7% | Step 5 Preview |
| DeepSearchQA | 93.1% | Coming soon | Coming soon |
| OSWorld-Verified | 83.4% | Coming soon | Coming soon |
| Finance Agent v2 | 53.9% | Coming soon | Coming soon |
| MCP Atlas | 82.2% | 85.6% | Step 5 Preview |
| Toolathlon | 59.9% | Coming soon | Coming soon |
| Gert Labs | 72.97% | Coming soon | Coming soon |
| ResearchClawBench | 21.1% | Coming soon | Coming soon |
| OSWorld 2.0 | 20.6% | Coming soon | Coming soon |
| Terminal-Bench 2.1 (Vals) | 71.9% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 85.0% | Coming soon |
| Terminal-Bench 4.0 | Coming soon | 33.30% | Coming soon |
| CyberGym | Coming soon | 84.7% | Coming soon |
| AutomationBench | Coming soon | 44.0% | Coming soon |
| Toolathlon-Verified | Coming soon | 74.1% | Coming soon |
| JobBench | Coming soon | 59.0% | Coming soon |
| APEX-Agents | Coming soon | 37.8% | Coming soon |
| DRACO | Coming soon | 83.3% | Coming soon |
| Agents' Last Exam | Coming soon | 29.5% | Coming soon |

### Coding

- Winner: Coming soon
- Claude Opus 4.8 public-lane score: 67.1 (Supported · #11/156)
- Step 5 Preview public-lane score: Coming soon

| Benchmark | Claude Opus 4.8 | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Verified | 88.6% | Coming soon | Coming soon |
| SWE-bench Pro | 69.2% | Coming soon | Coming soon |
| SWE Multilingual | 84.4% | Coming soon | Coming soon |
| SWE Multimodal | 38.4% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 74.6% | Coming soon | Coming soon |
| cursorBench31 | 58.4% | Coming soon | Coming soon |
| cursorBench32 | 62.3% | Coming soon | Coming soon |
| FrontierCode 1.1 Main | 46.5% | Coming soon | Coming soon |
| LiveCodeBench (Vals) | 87.8% | Coming soon | Coming soon |
| SWE-bench (Vals) | 88.6% | Coming soon | Coming soon |
| DeepSWE | Coming soon | 67.7% | 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: Coming soon
- Claude Opus 4.8 public-lane score: 55.5 (#17/19)
- Step 5 Preview public-lane score: 78.5 (Unranked · 2 rankable rows)

| Benchmark | Claude Opus 4.8 | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| ARC-AGI-2 | 72.1% | Coming soon | Coming soon |
| ARC-AGI-3 | 1.5% | Coming soon | Coming soon |
| CritPt | Coming soon | 20.9% | Coming soon |

### Multimodal & Grounded

- Winner: Coming soon
- Claude Opus 4.8 public-lane score: 87.7 (#3/49)
- Step 5 Preview public-lane score: 61.2 (#28/49)

| Benchmark | Claude Opus 4.8 | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| OfficeQA Pro | 66.2% | 60.3% | Claude Opus 4.8 |
| ScreenSpot Pro | 87.9% | Coming soon | Coming soon |
| CharXiv | 89.9% | Coming soon | Coming soon |
| CharXiv w/o tools | 80.5% | Coming soon | Coming soon |
| MMMU-Pro | Coming soon | 76% | Coming soon |

### Knowledge

- Winner: Coming soon
- Claude Opus 4.8 public-lane score: 71.7 (Supported · #9/186)
- Step 5 Preview public-lane score: Coming soon

| Benchmark | Claude Opus 4.8 | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| GPQA | 93.6% | Coming soon | Coming soon |
| GPQA-D | 93.6% | 93.5% | Claude Opus 4.8 |
| HLE | 57.9% | 46.5% | Claude Opus 4.8 |
| HLE w/o tools | 49.8% | Coming soon | Coming soon |
| GPQA Diamond (Vals) | 92.4% | Coming soon | Coming soon |
| MMLU-Pro (Vals) | 89.6% | Coming soon | Coming soon |

### Multilingual

- Winner: Coming soon
- Claude Opus 4.8 public-lane score: Coming soon
- Step 5 Preview public-lane score: Coming soon

| Benchmark | Claude Opus 4.8 | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| INCLUDE | 87.6% | Coming soon | Coming soon |

### Instruction Following

- Winner: Coming soon
- Claude Opus 4.8 public-lane score: 74 (#60/124)
- Step 5 Preview public-lane score: Coming soon

### Mathematics

- Winner: Coming soon
- Claude Opus 4.8 public-lane score: 65.2 (#2/7)
- Step 5 Preview public-lane score: Coming soon

| Benchmark | Claude Opus 4.8 | Step 5 Preview | Winner |
|-----------|-----------|-----------|--------|
| USAMO 2026 | 96.7% | Coming soon | Coming soon |
| FrontierMath v2 (Tiers 1-3) | 47.241% | Coming soon | Coming soon |
| FrontierMath v2 (Tier 4) | 31.250% | Coming soon | Coming soon |

## FAQ

### Which is better overall, Claude Opus 4.8 or Step 5 Preview?

Claude Opus 4.8 is ahead overall on BenchLM right now.

### Where is the biggest gap between Claude Opus 4.8 and Step 5 Preview?

The widest category gap is in multimodal & grounded, where the averages are 87.7 for Claude Opus 4.8 and 61.2 for Step 5 Preview.

### How many benchmarks does Claude Opus 4.8 cover on BenchLM?

Claude Opus 4.8 currently has 38 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

- [Claude Opus 4.8 vs Claude Fable 5.1](/compare/claude-fable-5-1-vs-claude-opus-4-8)
- [Step 5 Preview vs Claude Fable 5.1](/compare/claude-fable-5-1-vs-step-5-preview)
- [Claude Opus 4.8 vs GPT-6 Astra](/compare/claude-opus-4-8-vs-gpt-6-astra)
- [Step 5 Preview vs GPT-6 Astra](/compare/gpt-6-astra-vs-step-5-preview)

## Explore More

- [Claude Opus 4.8 profile](/models/claude-opus-4-8)
- [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)