- GPQA Diamond 93.5
- Best of HLE 46.5 / HLE w/ tools (text-only) 59.4 59.4
- AA-LCR v1.1 88.3
- CritPt 20.9
- DeepSWE v1.1 (SWE-agent) 67.7
- Terminal-Bench v2.1 85
- Terminal-Bench v4 33.3
- CyberGym 84.7
- SciCode 58.9
- RoadmapBench 54.3
- ProgramBench (Pass Rate) 80.5
- SWE-Marathon v1.1 (Partial Score) 72.7
- MLS-Bench-Lite 40.5
- Mean of SWE-Atlas-QnA 63.6 / SWE-Atlas-Test-writing 50.8 57.2
- StepCodeBench (avg@4, internal) 49
- GDPval-AA v2.1 (Artificial Analysis, 2026-09-20) 1566
- τ³-Banking 42.5
- AutomationBench (public); AutomationBench-AA 51.0 44
- AA-Briefcase v1.1 1433
- Toolathlon-Verified 74.1
- MCP-Atlas 85.6
- PresentBench 76.8
- OfficeQA Pro 60.3
- SpeadSheet v2 29.4
- JobBench 59
- Apex-Agents 37.8
- Draco 83.3
- BrowseComp 88.7
- Mean of FinStepBench-LiveSearch 74.5 / FinStepBench-CorporateValuation 60.6 / FinStepBench-FinanceDR 55.8 63.6
- FrontierFinance 66.4
- Agents Last Exam (ALE-CLI) 29.5
- MMMU-Pro 76
- GDP.pdf 14.8
Step 5 Preview
Step 5 Preview ranks first on 2 of the 58 benchmarks it has scores on.
StepFun Released 18 Sep 2026 65
Benchmarks with a score
2
Ranked first, out of five or more models
4
In the top three
StepFun Sep 2026 Step 5 Preview
Closed weights
No Epoch Capabilities Index score yet
- MMMU-Pro 76% #30 /72
- SciCode 58.9% #3 /47
- Harvey LAB (Legal Agent Benchmark) 9.17% #22 /80
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Humanity's Last Exam | 59.4% | #13 of 147 | Model card | |
| LMArena (Elo) | 1,448 | #64 of 333 | Official leaderboard | |
| AA-LCR 1.1 | 88.3% | #1 of 5 | Model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GPQA Diamond | 93.5% | #9 of 284 | Model card | |
| CritPt | 20.9% | #28 of 160 | Epoch AI | |
| ProofBench | 42% | #38 of 83 | Epoch AI | |
| ProofBench v1.1 | 42% | #34 of 51 | Official leaderboard | |
| MysteryMechanism | 18.5% | #26 of 29 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SciCode | 58.9% | #3 of 47 | Model card | |
| SWE-Marathon | 72.7% | #1 of 11 | Model card | |
| WebDev Arena | 1,570 | #26 of 105 | Epoch AI | |
| ProgramBench | 80.5% | #5 of 20 | Model card | |
| Code Arena | 1,571 | #37 of 124 | Official leaderboard | |
| Vibe Code Bench | 70.4% | #44 of 115 | Official leaderboard | |
| Code Migration | 39% | #32 of 78 | Official leaderboard | |
| DeepSWE | 67.7% | #21 of 51 | Model card | |
| Image-to-WebDev Arena | 1,550 | #26 of 55 | Official leaderboard | |
| SWE Atlas | 57.2% | #6 of 11 | Model card | |
| Code Migration v1.1 | 47% | #17 of 29 | Official leaderboard | |
| RSI-Exam | 35.2% | #13 of 18 | Official leaderboard | |
| MLS-Bench | 40.5% | #3 of 4 | Model card | |
| IOI v2 (Vals) | 38.1% | #38 of 45 | Official leaderboard | |
| FrontierCS 2 | 8.33% | #16 of 16 | Official leaderboard | |
| RoadmapBench | 54.3% | #1 of 1 | Model card | |
| StepCodeBench | 49% | #1 of 1 | Model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MCP Atlas | 85.6% | #4 of 40 | Model card | |
| BrowseComp | 88.7% | #11 of 77 | Model card | |
| τ³-bench Banking | 42.5% | #3 of 20 | Model card | |
| Terminal-Bench 2.1 | 85% | #15 of 53 | Model card | |
| AutomationBench | 44% | #15 of 41 | Model card | |
| DRACO | 83.3% | #6 of 11 | Model card | |
| Agent Arena | -0.03 | #35 of 53 | Official leaderboard | |
| Toolathlon-Verified | 74.1% | #9 of 13 | Model card | |
| Hermes Bench | 69.2% | #13 of 17 | Official leaderboard | |
| Terminal-Bench 4.0 | 33.3% | #18 of 23 | Model card |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Harvey LAB (Legal Agent Benchmark) | 9.17% | #22 of 80 | Official leaderboard | |
| APEX-Agents | 37.8% | #4 of 12 | Model card | |
| Legal Research Bench | 40.4% | #32 of 88 | Official leaderboard | |
| Tax Agent Bench | 27.3% | #35 of 79 | Official leaderboard | |
| Harvey LAB v1.1 (HLAB-Vals-1.1) | 12.9% | #15 of 33 | Official leaderboard | |
| Vals Legal Index | 70.2% | #15 of 33 | Official leaderboard | |
| Excel Modeling Benchmark v1.1 | 69.5% | #17 of 37 | Official leaderboard | |
| FrontierFinance | 66.4% | #1 of 2 | Model card | |
| Agents' Last Exam | 29.5% | #13 of 25 | Model card | |
| Vals Finance Agent v2 | 50.7% | #42 of 80 | Official leaderboard | |
| Excel Modeling Benchmark | 57.1% | #41 of 76 | Official leaderboard | |
| JobBench | 59% | #7 of 12 | Model card | |
| Vals Index v2.2 | 54.8% | #20 of 32 | Official leaderboard | |
| PresentBench | 76.8% | #2 of 3 | Model card | |
| Vals Finance Agent v2.1 | 58.9% | #24 of 35 | Official leaderboard | |
| Vals Finance Index | 55.4% | #25 of 35 | Official leaderboard | |
| Public Benefits Bench | 58.1% | #37 of 50 | Official leaderboard | |
| OfficeQA Pro | 60.3% | #11 of 14 | Model card | |
| GDPval-AA v2.1 | 1,566 | #7 of 8 | Model card | |
| AA-Briefcase v1.1 | 1,433 | #6 of 6 | Model card | |
| SpreadsheetBench 2 | 29.4% | #2 of 2 | Model card | |
| FinStepBench | 63.6% | #1 of 1 | Model card |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vision Arena | 1,247 | #48 of 136 | Official leaderboard | |
| MMMU-Pro | 76% | #30 of 72 | Model card | |
| GDP.pdf | 14.8% | #14 of 15 | Model card |
Safety and security
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveBench | 58.6% | #23 of 28 | Official leaderboard | |
| Vals Science Index | 29.6% | #19 of 23 | Official leaderboard | |
| Hermes Index | 32.8 | #15 of 17 | Official leaderboard |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.