Model comparison
Gemini 2.0 Flash-Lite vs Llama-3.3-70B-Instruct
Gemini 2.0 Flash-Lite is the stronger model overall, scoring 37.3 to 30.3 on the Noometry Index.
Last verified . 26 shared benchmarks.
Google
MetaSummary
- They share 26 benchmarks with published results for both. Gemini 2.0 Flash-Lite scores higher in 7 categories and Llama-3.3-70B-Instruct in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Gemini 2.0 Flash-Lite leads 31.8 to 14.3.
- The biggest single-benchmark swing is LiveBench Data Analysis: 65.5% for Gemini 2.0 Flash-Lite and 49.5% for Llama-3.3-70B-Instruct.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.
Side by side
| Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 37.3 | 30.3 |
| Released | 2025-02-05 | 2024-12-06 |
| Weights | Proprietary | Open |
| Context window | — | 128K |
| Max output | — | 4K |
| Input $ / M tokens | — | $0.10 |
| Output $ / M tokens | — | $0.32 |
| Results tracked | 32 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 2.0 Flash-Lite leads
Gemini 2.0 Flash-Lite: 37.4 (#181), Llama-3.3-70B-Instruct: 30.6 (#284)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Coding | 47.1% | 36.6% |
| LMArena Coding | 1322 | 1268 |
| SciCode | — | 26% |
| WeirdML | — | 14.4% |
| BigCodeBench Instruct | — | 46.9% |
| BigCodeBench Complete | — | 57.5% |
Agentic & Tool Use Not comparable
Gemini 2.0 Flash-Lite: —, Llama-3.3-70B-Instruct: 25.6 (#100)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 31.9% |
| BALROG | — | 23% |
Reasoning Gemini 2.0 Flash-Lite leads
Gemini 2.0 Flash-Lite: 21.2 (#203), Llama-3.3-70B-Instruct: 13.7 (#325)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Reasoning | 50.1% | 50.8% |
| LMArena Hard Prompts | 1324 | 1257 |
| DTBench | 52.5% | 59.5% |
| LiveBench Data Analysis | 65.5% | 49.5% |
| ForecastBench | 57.1 | 58.6 |
| LiveBench | 54.3% | 50.2% |
| SimpleBench | — | 19.9% |
| CritPt | — | 0% |
| LMCA | — | 17.5% |
| Epoch Capabilities Index | — | 127.32 |
Math Gemini 2.0 Flash-Lite leads
Gemini 2.0 Flash-Lite: 31.8 (#192), Llama-3.3-70B-Instruct: 14.3 (#292)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Math | 58.1% | 42.2% |
| LMArena Math | 1309 | 1268 |
| OTIS Mock AIME 2024-2025 | — | 5.1% |
| Omni-MATH | 37.4% | — |
| MATH Level 5 | — | 41.6% |
Knowledge Gemini 2.0 Flash-Lite leads
Gemini 2.0 Flash-Lite: 34.6 (#185), Llama-3.3-70B-Instruct: 30.5 (#220)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Expert | 1305 | 1225 |
| GPQA Diamond | — | 47.4% |
| MMLU-Pro | 72% | — |
| Confabulations | — | 22.8% |
| Vectara Hallucination Rate | — | 4.1% |
| GPQA (HELM) | 50% | — |
| MMLU | — | 86.3% |
Multimodal Not comparable
Gemini 2.0 Flash-Lite: 30.6 (#103), Llama-3.3-70B-Instruct: —
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Vision | 1100 | — |
Multilingual Gemini 2.0 Flash-Lite leads
Gemini 2.0 Flash-Lite: 46.2 (#156), Llama-3.3-70B-Instruct: 40.1 (#215)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | 1323 | 1236 |
| LMArena Chinese | 1339 | 1217 |
| LMArena French | 1346 | 1280 |
| LMArena German | 1306 | 1251 |
| LMArena Japanese | 1302 | 1149 |
| LMArena Korean | 1325 | 1142 |
| LMArena Russian | 1328 | 1252 |
| LMArena Spanish | 1313 | 1269 |
Instruction Following Too close to call
Gemini 2.0 Flash-Lite: 70.6 (#158), Llama-3.3-70B-Instruct: 71.3 (#152)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Instruction Following | 78.3% | 82.7% |
| LMArena Instruction Following | 1305 | 1242 |
| IFEval | 82.4% | — |
Long Context Gemini 2.0 Flash-Lite leads
Gemini 2.0 Flash-Lite: 40.1 (#155), Llama-3.3-70B-Instruct: 26.4 (#290)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Longer Query | 1321 | 1256 |
| Fiction.LiveBench | — | 33.3% |
Writing & Preference Gemini 2.0 Flash-Lite leads
Gemini 2.0 Flash-Lite: 51.9 (#172), Llama-3.3-70B-Instruct: 47.8 (#203)
| Benchmark | Gemini 2.0 Flash-Lite | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | 1330 | 1274 |
| LMArena Creative Writing | 1319 | 1250 |
| LMArena Multi-Turn | 1307 | 1279 |
| LiveBench Language | 34.3% | 39.2% |
| WildBench | 79% | — |
Frequently asked questions
Is Gemini 2.0 Flash-Lite better than Llama-3.3-70B-Instruct?
Gemini 2.0 Flash-Lite is the stronger model overall, scoring 37.3 to 30.3 on the Noometry Index.
Is Gemini 2.0 Flash-Lite or Llama-3.3-70B-Instruct better for coding?
Gemini 2.0 Flash-Lite scores higher on coding benchmarks: 37.4 versus 30.6 in the Noometry coding category.
How many benchmarks do Gemini 2.0 Flash-Lite and Llama-3.3-70B-Instruct share?
26 benchmarks have published results for both models. Gemini 2.0 Flash-Lite has 32 scored results on Noometry and Llama-3.3-70B-Instruct has 43.