Model comparison
DeepSeek-V2.5 (Sep 2024) vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 37.6 on the Noometry Index.
Last verified . 13 shared benchmarks.
DeepSeek
StepFunSummary
- They share 13 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 0 categories and Step 5 Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where Step 5 Preview leads 51.8 to 31.7.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Step 5 Preview | |
|---|---|---|
| Provider | DeepSeek | StepFun |
| Noometry Index | 37.6 | 47.9 |
| Released | 2024-09-06 | 2026-09-16 |
| Weights | Open | Proprietary |
| Context window | — | 1.02M |
| Max output | — | 66K |
| Input $ / M tokens | — | $1 |
| Output $ / M tokens | — | $2.70 |
| Results tracked | 22 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Step 5 Preview: 51.8 (#36)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Coding | 1309 | 1480 |
| Aider Polyglot | 17.8% | — |
| LMArena WebDev | — | 1564 |
| SciCode | — | 58.9% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Reasoning Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Step 5 Preview: 40.0 (#57)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1465 |
| CritPt | — | 20.9% |
Math Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Step 5 Preview: 46.1 (#72)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Math | 1288 | 1470 |
| ProofBench | — | 42% |
Knowledge Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Step 5 Preview: 41.2 (#112)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1266 | 1470 |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, Step 5 Preview: 41.0 (#41)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |
Multilingual Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Step 5 Preview: 53.8 (#54)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1273 | 1432 |
| LMArena Chinese | 1318 | 1519 |
| LMArena Russian | 1289 | 1436 |
| LMArena Spanish | 1248 | 1439 |
| LMArena French | 1289 | — |
| LMArena German | 1258 | — |
| LMArena Japanese | 1228 | — |
| LMArena Korean | 1209 | — |
Instruction Following Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Step 5 Preview: 76.0 (#50)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1280 | 1444 |
Long Context Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Step 5 Preview: 44.5 (#64)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1301 | 1455 |
Writing & Preference Step 5 Preview leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Step 5 Preview: 62.8 (#71)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Step 5 Preview |
|---|---|---|
| LMArena Text | 1294 | 1442 |
| LMArena Creative Writing | 1285 | 1410 |
| LMArena Multi-Turn | 1297 | 1447 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Step 5 Preview share?
13 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Step 5 Preview has 18.