Model comparison
Amazon Nova Lite vs Llama-3.3-70B-Instruct
Amazon Nova Lite is the stronger model overall, scoring 31.9 to 30.6 on the Noometry Index.
Last verified . 26 shared benchmarks.
MetaSummary
- They share 26 benchmarks with published results for both. Amazon Nova Lite scores higher in 4 categories and Llama-3.3-70B-Instruct in 4 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Amazon Nova Lite leads 27.8 to 15.3.
- The biggest single-benchmark swing is LiveBench Instruction Following: 54.1% for Amazon Nova Lite and 82.7% for Llama-3.3-70B-Instruct.
- Amazon Nova Lite is cheaper at $0.06 / $0.24 per million input/output tokens, against $0.10 / $0.32 for Llama-3.3-70B-Instruct.
- Amazon Nova Lite accepts more context: 300K tokens versus 128K.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.
Side by side
| Amazon Nova Lite | Llama-3.3-70B-Instruct | |
|---|---|---|
| Provider | Amazon | Meta |
| Noometry Index | 31.9 | 30.6 |
| Released | 2024-12-03 | 2024-12-06 |
| Weights | Proprietary | Open |
| Context window | 300K | 128K |
| Max output | 10K | 4K |
| Input $ / M tokens | $0.06 | $0.10 |
| Output $ / M tokens | $0.24 | $0.32 |
| Results tracked | 34 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Amazon Nova Lite leads
Amazon Nova Lite: 32.5 (#270), Llama-3.3-70B-Instruct: 31.0 (#290)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Coding | 27.5% | 36.6% |
| LMArena Coding | 1239 | 1268 |
| SciCode | — | 26% |
| WeirdML | — | 14.4% |
| BigCodeBench Instruct | — | 46.9% |
| BigCodeBench Complete | — | 57.5% |
| ALE-Bench | 236.25 | — |
Agentic & Tool Use Not comparable
Amazon Nova Lite: —, Llama-3.3-70B-Instruct: 25.8 (#105)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 31.9% |
| BALROG | — | 23% |
Reasoning Amazon Nova Lite leads
Amazon Nova Lite: 19.3 (#260), Llama-3.3-70B-Instruct: 14.1 (#327)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Reasoning | 36.7% | 50.8% |
| LMArena Hard Prompts | 1220 | 1257 |
| LiveBench Data Analysis | 37.2% | 49.5% |
| LiveBench | 36.4% | 50.2% |
| SimpleBench | — | 19.9% |
| CritPt | — | 0% |
| DTBench | — | 59.5% |
| LMCA | — | 17.5% |
| Epoch Capabilities Index | — | 127.33 |
| ForecastBench | — | 58.6 |
Math Amazon Nova Lite leads
Amazon Nova Lite: 27.8 (#247), Llama-3.3-70B-Instruct: 15.3 (#298)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Math | 36.7% | 42.2% |
| LMArena Math | 1227 | 1267 |
| OTIS Mock AIME 2024-2025 | — | 5.1% |
| Omni-MATH | 23.3% | — |
| MATH Level 5 | — | 41.6% |
Knowledge Llama-3.3-70B-Instruct leads
Amazon Nova Lite: 26.3 (#259), Llama-3.3-70B-Instruct: 30.6 (#226)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| Vectara Hallucination Rate | 6.1% | 4.1% |
| LMArena Expert | 1201 | 1225 |
| MMLU | 77% | 86.3% |
| GPQA Diamond | — | 47.4% |
| Humanity's Last Exam | 3.6% | — |
| MMLU-Pro | 60% | — |
| Confabulations | — | 22.8% |
| GPQA (HELM) | 39.7% | — |
Multimodal Not comparable
Amazon Nova Lite: 25.5 (#123), Llama-3.3-70B-Instruct: —
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Vision | 990 | — |
Multilingual Llama-3.3-70B-Instruct leads
Amazon Nova Lite: 38.0 (#232), Llama-3.3-70B-Instruct: 39.9 (#220)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | 1208 | 1236 |
| LMArena Chinese | 1225 | 1217 |
| LMArena French | 1237 | 1281 |
| LMArena German | 1229 | 1251 |
| LMArena Japanese | 1153 | 1150 |
| LMArena Korean | 1153 | 1143 |
| LMArena Russian | 1216 | 1252 |
| LMArena Spanish | 1226 | 1270 |
Instruction Following Llama-3.3-70B-Instruct leads
Amazon Nova Lite: 59.3 (#255), Llama-3.3-70B-Instruct: 71.1 (#157)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Instruction Following | 54.1% | 82.7% |
| LMArena Instruction Following | 1205 | 1242 |
| IFEval | 77.6% | — |
Long Context Amazon Nova Lite leads
Amazon Nova Lite: 37.4 (#217), Llama-3.3-70B-Instruct: 26.4 (#295)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Longer Query | 1234 | 1256 |
| Fiction.LiveBench | — | 33.3% |
Writing & Preference Llama-3.3-70B-Instruct leads
Amazon Nova Lite: 42.1 (#237), Llama-3.3-70B-Instruct: 47.6 (#207)
| Benchmark | Amazon Nova Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | 1229 | 1274 |
| LMArena Creative Writing | 1197 | 1250 |
| LMArena Multi-Turn | 1199 | 1280 |
| LiveBench Language | 25.9% | 39.2% |
| WildBench | 75% | — |
Frequently asked questions
Is Amazon Nova Lite better than Llama-3.3-70B-Instruct?
Amazon Nova Lite is the stronger model overall, scoring 31.9 to 30.6 on the Noometry Index.
Which is cheaper, Amazon Nova Lite or Llama-3.3-70B-Instruct?
Amazon Nova Lite is cheaper. It lists at $0.06 per million input tokens and $0.24 per million output tokens; Llama-3.3-70B-Instruct lists at $0.10 and $0.32.
Is Amazon Nova Lite or Llama-3.3-70B-Instruct better for coding?
Amazon Nova Lite scores higher on coding benchmarks: 32.5 versus 31.0 in the Noometry coding category.
Which has the bigger context window?
Amazon Nova Lite does, with 300K tokens against 128K.
How many benchmarks do Amazon Nova Lite and Llama-3.3-70B-Instruct share?
26 benchmarks have published results for both models. Amazon Nova Lite has 34 scored results on Noometry and Llama-3.3-70B-Instruct has 43.