Model comparison
Llama-3.3-70B-Instruct vs Step 5 Preview
Step 5 Preview is the stronger model overall, scoring 47.9 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 9.2× 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.
Meta
StepFunSummary
- They share 15 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 0 categories and Step 5 Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Step 5 Preview leads 46.1 to 15.3.
- The biggest single-benchmark swing is SciCode: 26% for Llama-3.3-70B-Instruct and 58.9% for Step 5 Preview.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $1 / $2.70 for Step 5 Preview.
- Step 5 Preview accepts more context: 1.02M tokens versus 128K.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.
Side by side
| Llama-3.3-70B-Instruct | Step 5 Preview | |
|---|---|---|
| Provider | Meta | StepFun |
| Noometry Index | 30.6 | 47.9 |
| Released | 2024-12-06 | 2026-09-16 |
| Weights | Open | Proprietary |
| Context window | 128K | 1.02M |
| Max output | 4K | 66K |
| Input $ / M tokens | $0.10 | $1 |
| Output $ / M tokens | $0.32 | $2.70 |
| Results tracked | 43 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Step 5 Preview leads
Llama-3.3-70B-Instruct: 31.0 (#290), Step 5 Preview: 51.8 (#36)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| SciCode | 26% | 58.9% |
| LMArena Coding | 1268 | 1480 |
| LMArena WebDev | — | 1564 |
| WeirdML | 14.4% | — |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
| BigCodeBench Complete | 57.5% | — |
Agentic & Tool Use Not comparable
Llama-3.3-70B-Instruct: 25.8 (#105), Step 5 Preview: —
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | — |
| BALROG | 23% | — |
Reasoning Step 5 Preview leads
Llama-3.3-70B-Instruct: 14.1 (#327), Step 5 Preview: 40.0 (#57)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| CritPt | 0% | 20.9% |
| LMArena Hard Prompts | 1257 | 1465 |
| SimpleBench | 19.9% | — |
| LiveBench Reasoning | 50.8% | — |
| DTBench | 59.5% | — |
| LiveBench Data Analysis | 49.5% | — |
| LMCA | 17.5% | — |
| Epoch Capabilities Index | 127.33 | — |
| ForecastBench | 58.6 | — |
| LiveBench | 50.2% | — |
Math Step 5 Preview leads
Llama-3.3-70B-Instruct: 15.3 (#298), Step 5 Preview: 46.1 (#72)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| LMArena Math | 1267 | 1470 |
| OTIS Mock AIME 2024-2025 | 5.1% | — |
| ProofBench | — | 42% |
| LiveBench Math | 42.2% | — |
| MATH Level 5 | 41.6% | — |
Knowledge Step 5 Preview leads
Llama-3.3-70B-Instruct: 30.6 (#226), Step 5 Preview: 41.2 (#112)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1225 | 1470 |
| GPQA Diamond | 47.4% | — |
| Confabulations | 22.8% | — |
| Vectara Hallucination Rate | 4.1% | — |
| MMLU | 86.3% | — |
Multimodal Not comparable
Llama-3.3-70B-Instruct: —, Step 5 Preview: 41.0 (#41)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |
Multilingual Step 5 Preview leads
Llama-3.3-70B-Instruct: 39.9 (#220), Step 5 Preview: 53.8 (#54)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1236 | 1432 |
| LMArena Chinese | 1217 | 1519 |
| LMArena Russian | 1252 | 1436 |
| LMArena Spanish | 1270 | 1439 |
| LMArena French | 1281 | — |
| LMArena German | 1251 | — |
| LMArena Japanese | 1150 | — |
| LMArena Korean | 1143 | — |
Instruction Following Step 5 Preview leads
Llama-3.3-70B-Instruct: 71.1 (#157), Step 5 Preview: 76.0 (#50)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1242 | 1444 |
| LiveBench Instruction Following | 82.7% | — |
Long Context Step 5 Preview leads
Llama-3.3-70B-Instruct: 26.4 (#295), Step 5 Preview: 44.5 (#64)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1256 | 1455 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Step 5 Preview leads
Llama-3.3-70B-Instruct: 47.6 (#207), Step 5 Preview: 62.8 (#71)
| Benchmark | Llama-3.3-70B-Instruct | Step 5 Preview |
|---|---|---|
| LMArena Text | 1274 | 1442 |
| LMArena Creative Writing | 1250 | 1410 |
| LMArena Multi-Turn | 1280 | 1447 |
| LiveBench Language | 39.2% | — |
Frequently asked questions
Is Llama-3.3-70B-Instruct better than Step 5 Preview?
Step 5 Preview is the stronger model overall, scoring 47.9 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 9.2× less per token, which makes it the better buy when Step 5 Preview's lead doesn't matter for your workload.
Which is cheaper, Llama-3.3-70B-Instruct or Step 5 Preview?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; Step 5 Preview lists at $1 and $2.70.
Is Llama-3.3-70B-Instruct or Step 5 Preview better for coding?
Step 5 Preview scores higher on coding benchmarks: 51.8 versus 31.0 in the Noometry coding category.
Which has the bigger context window?
Step 5 Preview does, with 1.02M tokens against 128K.
How many benchmarks do Llama-3.3-70B-Instruct and Step 5 Preview share?
15 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and Step 5 Preview has 18.