Model comparison
DeepSeek-R1 vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 42.3 on the Noometry Index. DeepSeek-R1 costs 1.6× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Last verified . 15 shared benchmarks.
DeepSeek
StepFunSummary
- They share 15 benchmarks with published results for both. DeepSeek-R1 scores higher in 2 categories and Step 5 Preview in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Step 5 Preview leads 40.0 to 18.6.
- The biggest single-benchmark swing is SciCode: 35.7% for DeepSeek-R1 and 58.9% for Step 5 Preview.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $1 / $2.70 for Step 5 Preview.
- Step 5 Preview accepts more context: 1.02M tokens versus 164K.
Side by side
| DeepSeek-R1 | Step 5 Preview | |
|---|---|---|
| Provider | DeepSeek | StepFun |
| Noometry Index | 42.3 | 47.9 |
| Released | 2025-01-20 | 2026-09-16 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 1.02M |
| Max output | 64K | 66K |
| Input $ / M tokens | $0.50 | $1 |
| Output $ / M tokens | $2.15 | $2.70 |
| Results tracked | 52 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
DeepSeek-R1: 46.3 (#68), Step 5 Preview: 51.8 (#36)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| SciCode | 35.7% | 58.9% |
| LMArena Coding | 1427 | 1480 |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1564 |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Not comparable
DeepSeek-R1: 30.7 (#75), Step 5 Preview: —
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning Step 5 Preview leads
DeepSeek-R1: 18.6 (#278), Step 5 Preview: 40.0 (#57)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| CritPt | 1.1% | 20.9% |
| LMArena Hard Prompts | 1416 | 1465 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| ARC-AGI-1 | 21.2% | — |
| LiveBench Reasoning | 83.2% | — |
| LiveBench Data Analysis | 69.8% | — |
| Epoch Capabilities Index | 141.29 | — |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math Step 5 Preview leads
DeepSeek-R1: 43.8 (#79), Step 5 Preview: 46.1 (#72)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| LMArena Math | 1400 | 1470 |
| OTIS Mock AIME 2024-2025 | 66.4% | — |
| ProofBench | — | 42% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Step 5 Preview: 41.2 (#112)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1394 | 1470 |
| GPQA Diamond | 76.3% | — |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, Step 5 Preview: 41.0 (#41)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |
Multilingual Step 5 Preview leads
DeepSeek-R1: 52.4 (#85), Step 5 Preview: 53.8 (#54)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1412 | 1432 |
| LMArena Chinese | 1442 | 1519 |
| LMArena Russian | 1423 | 1436 |
| LMArena Spanish | 1411 | 1439 |
| LMArena French | 1417 | — |
| LMArena German | 1404 | — |
| LMArena Japanese | 1391 | — |
| LMArena Korean | 1360 | — |
Instruction Following Step 5 Preview leads
DeepSeek-R1: 72.0 (#143), Step 5 Preview: 76.0 (#50)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1382 | 1444 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context Too close to call
DeepSeek-R1: 45.4 (#36), Step 5 Preview: 44.5 (#64)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1391 | 1455 |
| Fiction.LiveBench | 75% | — |
Writing & Preference Step 5 Preview leads
DeepSeek-R1: 61.4 (#88), Step 5 Preview: 62.8 (#71)
| Benchmark | DeepSeek-R1 | Step 5 Preview |
|---|---|---|
| LMArena Text | 1428 | 1442 |
| LMArena Creative Writing | 1405 | 1410 |
| LMArena Multi-Turn | 1405 | 1447 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 42.3 on the Noometry Index. DeepSeek-R1 costs 1.6× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or Step 5 Preview?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Step 5 Preview lists at $1 and $2.70.
Is DeepSeek-R1 or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
Step 5 Preview does, with 1.02M tokens against 164K.
How many benchmarks do DeepSeek-R1 and Step 5 Preview share?
15 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Step 5 Preview has 18.