# Solar Pro 4 Benchmark Scores & Performance

> Solar Pro 4 by Upstage has 22 source-displayable benchmark rows but no public overall score. It remains unranked.

Canonical page: https://benchlm.ai/models/solar-pro-4

Last updated: September 10, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Upstage |
| Source Type | Proprietary |
| Reasoning Type | Reasoning |
| Context Window | 512K |
| Official model card | [Upstage Solar Pro 4 launch post](https://www.upstage.ai/blog/en/solar-pro-4) |
| Overall Score | Coming soon |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Solar
- Variant: pro
- Benchmarks covered: 22 of 434
- Sibling models: [Solar Pro 2](/models/solar-pro-2), [Solar Pro 3](/models/solar-pro-3)
- Related earlier model: [Solar Pro 3](/models/solar-pro-3)
- Coverage note: BenchLM currently has partial benchmark coverage for this model, so the overall score is conservative.

## Agentic Benchmarks

| Benchmark | Score |
|-----------|-------|
| [Terminal-Bench 2.1](/benchmarks/terminalbench21) | 57.0% |
| [BrowseComp](/benchmarks/browsecomp) | 49.2% |
| [MCP Atlas](/benchmarks/mcpatlas) | 61.4% |
| [APEX-Agents](/benchmarks/apexagents) | 18.7% |
| [GDPval-AA](/benchmarks/gdpvalaanormalized) | 34.7% |

## Coding Benchmarks

| Benchmark | Score |
|-----------|-------|
| [Terminal-Bench 2.1](/benchmarks/terminalbench21) | 57.0% |
| [SWE-bench Verified](/benchmarks/swe-bench-verified) | 70.6% |
| [LiveCodeBench](/benchmarks/livecodebench) | 87.8% |
| [AA-SciCode](/benchmarks/aascicode) | 44.6% |

## Multimodal & Grounded Benchmarks

| Benchmark | Score |
|-----------|-------|
| [Design Arena Website](/benchmarks/designarenawebsite) | 1188 |

## Reasoning Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AA-LCR](/benchmarks/lcr) | 71.0% |
| [CritPt](/benchmarks/critpt) | 5.4% |

## Knowledge Benchmarks

| Benchmark | Score |
|-----------|-------|
| [GPQA-D](/benchmarks/gpqa-diamond) | 89.0% |
| [MMLU-Pro](/benchmarks/mmlu-pro) | 86.3% |
| [Artificial Analysis Intelligence Index](/benchmarks/artificialanalysis) | 28.1% |
| [AA-GPQA Diamond](/benchmarks/aagpqadiamond) | 89.1% |
| [AA-HLE](/benchmarks/aahle) | 29.2% |
| [AA-Omniscience Index](/benchmarks/aaomniscienceindex) | -0.8% |
| [AA-Omniscience Accuracy](/benchmarks/omniscienceaccuracy) | 18.9% |
| [AA-Omniscience Hallucination Rate](/benchmarks/omnisciencehallucinationrate) | 24.4% |

## Korean Benchmarks Benchmarks

| Benchmark | Score |
|-----------|-------|
| [KMMLU-Pro](/benchmarks/kmmlupro) | 79.2% |

## Mathematics Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AIME26](/benchmarks/aime2026) | 95.3% |

## Other Upstage Models

- [Solar Pro 2](/models/solar-pro-2) - Score: 38.86
- [Solar Pro 3](/models/solar-pro-3) - Score: not computed
- [Solar Open 2](/models/solar-open-2) - Score: not computed