Model comparison
Llama-3.3-70B-Instruct vs Llama 4 Scout
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.3 to 27.4 on the Noometry Index.
Last verified . 30 shared benchmarks.
MetaSummary
- They share 30 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 5 categories and Llama 4 Scout in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in coding, where Llama-3.3-70B-Instruct leads 30.6 to 19.9.
- The biggest single-benchmark swing is MATH Level 5: 41.6% for Llama-3.3-70B-Instruct and 62.3% for Llama 4 Scout.
- Both cost about the same: $0.10 input and $0.32 output per million tokens.
Side by side
| Llama-3.3-70B-Instruct | Llama 4 Scout | |
|---|---|---|
| Provider | Meta | Meta |
| Noometry Index | 30.3 | 27.4 |
| Released | 2024-12-06 | 2025-04-05 |
| Weights | Open | Open |
| Context window | 128K | 128K |
| Max output | 4K | 4K |
| Input $ / M tokens | $0.10 | $0.10 |
| Output $ / M tokens | $0.32 | $0.30 |
| Results tracked | 43 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 30.6 (#284), Llama 4 Scout: 19.9 (#334)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| SciCode | 26% | 17% |
| LMArena Coding | 1268 | 1286 |
| BigCodeBench Complete | 57.5% | 43.1% |
| SWE-bench Verified (bash only) | — | 9.1% |
| WeirdML | 14.4% | — |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
Agentic & Tool Use Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 25.6 (#100), Llama 4 Scout: 24.4 (#109)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | 28.1% |
| BALROG | 23% | — |
Reasoning Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 13.7 (#325), Llama 4 Scout: 9.2 (#339)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| CritPt | 0% | 0% |
| LMArena Hard Prompts | 1257 | 1266 |
| DTBench | 59.5% | 57.9% |
| LMCA | 17.5% | 12% |
| Epoch Capabilities Index | 127.32 | 129.64 |
| ForecastBench | 58.6 | 57.5 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | 19.9% | — |
| Kagi LLM Benchmark | — | 36.9% |
| ARC-AGI-1 | — | 0.5% |
| LiveBench Reasoning | 50.8% | — |
| LiveBench Data Analysis | 49.5% | — |
| LiveBench | 50.2% | — |
Math Llama 4 Scout leads
Llama-3.3-70B-Instruct: 14.3 (#292), Llama 4 Scout: 18.4 (#280)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 5.1% | 7.8% |
| LMArena Math | 1268 | 1286 |
| MATH Level 5 | 41.6% | 62.3% |
| Omni-MATH | — | 37.3% |
| LiveBench Math | 42.2% | — |
| FrontierMath (Feb 2025 set) | — | 0% |
Knowledge Llama 4 Scout leads
Llama-3.3-70B-Instruct: 30.5 (#220), Llama 4 Scout: 31.8 (#208)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| GPQA Diamond | 47.4% | 51.8% |
| Vectara Hallucination Rate | 4.1% | 7.7% |
| LMArena Expert | 1225 | 1236 |
| MMLU-Pro | — | 74.2% |
| Confabulations | 22.8% | — |
| GPQA (HELM) | — | 50.7% |
| MMLU | 86.3% | — |
Multimodal Not comparable
Llama-3.3-70B-Instruct: —, Llama 4 Scout: 31.6 (#98)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| LMArena Vision | — | 1118 |
| SpatialViz-Bench | — | 34.2% |
Multilingual Llama 4 Scout leads
Llama-3.3-70B-Instruct: 40.1 (#215), Llama 4 Scout: 41.3 (#207)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| LMArena Non-English | 1236 | 1253 |
| LMArena Chinese | 1217 | 1256 |
| LMArena French | 1280 | 1283 |
| LMArena German | 1251 | 1274 |
| LMArena Japanese | 1149 | 1209 |
| LMArena Korean | 1142 | 1207 |
| LMArena Russian | 1252 | 1263 |
| LMArena Spanish | 1269 | 1277 |
Instruction Following Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 71.3 (#152), Llama 4 Scout: 66.0 (#212)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| LMArena Instruction Following | 1242 | 1248 |
| LiveBench Instruction Following | 82.7% | — |
| IFEval | — | 81.8% |
Long Context Llama 4 Scout leads
Llama-3.3-70B-Instruct: 26.4 (#290), Llama 4 Scout: 27.5 (#289)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| Fiction.LiveBench | 33.3% | 36% |
| LMArena Longer Query | 1256 | 1265 |
Writing & Preference Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 47.8 (#203), Llama 4 Scout: 37.1 (#256)
| Benchmark | Llama-3.3-70B-Instruct | Llama 4 Scout |
|---|---|---|
| LMArena Text | 1274 | 1279 |
| LMArena Creative Writing | 1250 | 1249 |
| LMArena Multi-Turn | 1279 | 1281 |
| EQ-Bench Creative Writing | — | 783 |
| WildBench | — | 78% |
| LiveBench Language | 39.2% | — |
Frequently asked questions
Is Llama-3.3-70B-Instruct better than Llama 4 Scout?
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.3 to 27.4 on the Noometry Index.
Which is cheaper, Llama-3.3-70B-Instruct or Llama 4 Scout?
Llama 4 Scout is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; Llama-3.3-70B-Instruct lists at $0.10 and $0.32.
Is Llama-3.3-70B-Instruct or Llama 4 Scout better for coding?
Llama-3.3-70B-Instruct scores higher on coding benchmarks: 30.6 versus 19.9 in the Noometry coding category.
Which has the bigger context window?
Both accept 128K tokens.
How many benchmarks do Llama-3.3-70B-Instruct and Llama 4 Scout share?
30 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and Llama 4 Scout has 43.