Model comparison
GPT-3.5-turbo vs Llama-3.3-70B-Instruct
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.6 to 23.2 on the Noometry Index.
Last verified . 28 shared benchmarks.
MetaSummary
- They share 28 benchmarks with published results for both. GPT-3.5-turbo scores higher in 1 category and Llama-3.3-70B-Instruct in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Llama-3.3-70B-Instruct leads 47.6 to 25.3.
- The biggest single-benchmark swing is MATH Level 5: 15.9% for GPT-3.5-turbo and 41.6% for Llama-3.3-70B-Instruct.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $0.50 / $1.50 for GPT-3.5-turbo.
- Llama-3.3-70B-Instruct accepts more context: 128K tokens versus 16K.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-3.5-turbo | Llama-3.3-70B-Instruct | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 23.2 | 30.6 |
| Released | 2023-03-01 | 2024-12-06 |
| Weights | Proprietary | Open |
| Context window | 16K | 128K |
| Max output | 4K | 4K |
| Input $ / M tokens | $0.50 | $0.10 |
| Output $ / M tokens | $1.50 | $0.32 |
| Results tracked | 44 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama-3.3-70B-Instruct leads
GPT-3.5-turbo: 23.9 (#331), Llama-3.3-70B-Instruct: 31.0 (#290)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| WeirdML | 3.5% | 14.4% |
| BigCodeBench Instruct | 39.1% | 46.9% |
| LMArena Coding | 1136 | 1268 |
| BigCodeBench Complete | 50.6% | 57.5% |
| SciCode | — | 26% |
| LiveBench Coding | — | 36.6% |
| HumanEval+ | 70.7% | — |
| MBPP+ | 69.7% | — |
Agentic & Tool Use Not comparable
GPT-3.5-turbo: —, Llama-3.3-70B-Instruct: 25.8 (#105)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 31.9% |
| BALROG | — | 23% |
| METR Time Horizons | 21.5% | — |
Reasoning Too close to call
GPT-3.5-turbo: 13.8 (#332), Llama-3.3-70B-Instruct: 14.1 (#327)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Hard Prompts | 1108 | 1257 |
| DTBench | 48.5% | 59.5% |
| LMCA | 9.7% | 17.5% |
| Epoch Capabilities Index | 118.55 | 127.33 |
| ForecastBench | 50.4 | 58.6 |
| SimpleBench | — | 19.9% |
| CritPt | — | 0% |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | — | 50.8% |
| Mystery Game Puzzles | 3% | — |
| LiveBench Data Analysis | — | 49.5% |
| Adversarial NLI | 58.1% | — |
| BIG-Bench Hard | 61.6% | — |
| CommonsenseQA 2.0 | 57% | — |
| LiveBench | — | 50.2% |
| WinoGrande | 81.6% | — |
Math Llama-3.3-70B-Instruct leads
GPT-3.5-turbo: 6.3 (#327), Llama-3.3-70B-Instruct: 15.3 (#298)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.2% | 5.1% |
| LMArena Math | 1142 | 1267 |
| MATH Level 5 | 15.9% | 41.6% |
| FrontierMath (Tiers 1-3) | 0% | — |
| LiveBench Math | — | 42.2% |
| GSM8K | 57.8% | — |
Knowledge Llama-3.3-70B-Instruct leads
GPT-3.5-turbo: 10.0 (#303), Llama-3.3-70B-Instruct: 30.6 (#226)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| GPQA Diamond | 28% | 47.4% |
| LMArena Expert | 1070 | 1225 |
| MMLU | 71.4% | 86.3% |
| Confabulations | — | 22.8% |
| Vectara Hallucination Rate | — | 4.1% |
| ARC (AI2) Challenge | 87.4% | — |
| BoolQ | 87% | — |
| OpenBookQA | 86% | — |
| TriviaQA | 85.8% | — |
Multilingual Llama-3.3-70B-Instruct leads
GPT-3.5-turbo: 31.5 (#258), Llama-3.3-70B-Instruct: 39.9 (#220)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | 1108 | 1236 |
| LMArena Chinese | 1075 | 1217 |
| LMArena French | 1118 | 1281 |
| LMArena German | 1090 | 1251 |
| LMArena Japanese | 1043 | 1150 |
| LMArena Korean | 1019 | 1143 |
| LMArena Russian | 1123 | 1252 |
| LMArena Spanish | 1121 | 1270 |
Instruction Following Llama-3.3-70B-Instruct leads
GPT-3.5-turbo: 57.9 (#262), Llama-3.3-70B-Instruct: 71.1 (#157)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Instruction Following | 1119 | 1242 |
| LiveBench Instruction Following | — | 82.7% |
Long Context GPT-3.5-turbo leads
GPT-3.5-turbo: 34.0 (#254), Llama-3.3-70B-Instruct: 26.4 (#295)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Longer Query | 1121 | 1256 |
| Fiction.LiveBench | — | 33.3% |
Writing & Preference Llama-3.3-70B-Instruct leads
GPT-3.5-turbo: 25.3 (#305), Llama-3.3-70B-Instruct: 47.6 (#207)
| Benchmark | GPT-3.5-turbo | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | 1125 | 1274 |
| LMArena Creative Writing | 1092 | 1250 |
| LMArena Multi-Turn | 1117 | 1280 |
| EQ-Bench Creative Writing | 451 | — |
| LiveBench Language | — | 39.2% |
Frequently asked questions
Is GPT-3.5-turbo better than Llama-3.3-70B-Instruct?
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.6 to 23.2 on the Noometry Index.
Which is cheaper, GPT-3.5-turbo or Llama-3.3-70B-Instruct?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; GPT-3.5-turbo lists at $0.50 and $1.50.
Is GPT-3.5-turbo or Llama-3.3-70B-Instruct better for coding?
Llama-3.3-70B-Instruct scores higher on coding benchmarks: 31.0 versus 23.9 in the Noometry coding category.
Which has the bigger context window?
Llama-3.3-70B-Instruct does, with 128K tokens against 16K.
How many benchmarks do GPT-3.5-turbo and Llama-3.3-70B-Instruct share?
28 benchmarks have published results for both models. GPT-3.5-turbo has 44 scored results on Noometry and Llama-3.3-70B-Instruct has 43.