Model comparison
DeepSeek-V3.1 vs Step 2 16k Exp 202412
DeepSeek-V3.1 is the stronger model overall, scoring 42.8 to 39.2 on the Noometry Index.
Last verified . 12 shared benchmarks.
DeepSeek
StepFunSummary
- They share 12 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 7 categories and Step 2 16k Exp 202412 in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.1 leads 43.7 to 35.2.
- DeepSeek-V3.1 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.1 | Step 2 16k Exp 202412 | |
|---|---|---|
| Provider | DeepSeek | StepFun |
| Noometry Index | 42.8 | 39.2 |
| Released | 2025-08-21 | — |
| Weights | Open | Proprietary |
| Context window | 164K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.25 | — |
| Output $ / M tokens | $0.95 | — |
| Results tracked | 27 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.1 leads
DeepSeek-V3.1: 40.3 (#144), Step 2 16k Exp 202412: 38.5 (#174)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Coding | 1417 | 1316 |
| WeirdML | 38.4% | — |
Reasoning DeepSeek-V3.1 leads
DeepSeek-V3.1: 27.9 (#110), Step 2 16k Exp 202412: 25.9 (#142)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Hard Prompts | 1417 | 1299 |
| SimpleBench | 40% | — |
| Kagi LLM Benchmark | 53.2% | — |
| DTBench | 82.7% | — |
| LMCA | 24.3% | — |
| Epoch Capabilities Index | 139.92 | — |
| ForecastBench | 58 | — |
Math DeepSeek-V3.1 leads
DeepSeek-V3.1: 38.9 (#122), Step 2 16k Exp 202412: 36.3 (#169)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Math | 1420 | 1304 |
Knowledge DeepSeek-V3.1 leads
DeepSeek-V3.1: 43.7 (#90), Step 2 16k Exp 202412: 35.2 (#187)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Expert | 1405 | 1279 |
| Vectara Hallucination Rate | 5.5% | — |
Multilingual DeepSeek-V3.1 leads
DeepSeek-V3.1: 51.6 (#106), Step 2 16k Exp 202412: 43.7 (#181)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Non-English | 1400 | 1290 |
| LMArena Chinese | 1469 | 1331 |
| LMArena Russian | 1405 | 1324 |
| LMArena French | 1447 | — |
| LMArena German | 1411 | — |
| LMArena Japanese | 1378 | — |
| LMArena Korean | 1337 | — |
| LMArena Spanish | 1431 | — |
Instruction Following DeepSeek-V3.1 leads
DeepSeek-V3.1: 73.9 (#110), Step 2 16k Exp 202412: 67.9 (#192)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Instruction Following | 1400 | 1287 |
Long Context Step 2 16k Exp 202412 leads
DeepSeek-V3.1: 36.3 (#232), Step 2 16k Exp 202412: 39.7 (#170)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Longer Query | 1422 | 1307 |
| Fiction.LiveBench | 52.8% | — |
Writing & Preference DeepSeek-V3.1 leads
DeepSeek-V3.1: 60.3 (#98), Step 2 16k Exp 202412: 52.3 (#173)
| Benchmark | DeepSeek-V3.1 | Step 2 16k Exp 202412 |
|---|---|---|
| LMArena Text | 1420 | 1321 |
| LMArena Creative Writing | 1401 | 1328 |
| LMArena Multi-Turn | 1408 | 1292 |
| EQ-Bench Creative Writing | 1436 | — |
Frequently asked questions
Is DeepSeek-V3.1 better than Step 2 16k Exp 202412?
DeepSeek-V3.1 is the stronger model overall, scoring 42.8 to 39.2 on the Noometry Index.
Is DeepSeek-V3.1 or Step 2 16k Exp 202412 better for coding?
DeepSeek-V3.1 scores higher on coding benchmarks: 40.3 versus 38.5 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.1 and Step 2 16k Exp 202412 share?
12 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Step 2 16k Exp 202412 has 12.