Model comparison
GPT-3.5-turbo vs Llama 3.1-8B
GPT-3.5-turbo and Llama 3.1-8B score almost the same on the Noometry Index (23.2 vs 23.0), so choose on price, context window or the category you care about most.
Last verified . 33 shared benchmarks.
MetaSummary
- They share 33 benchmarks with published results for both. GPT-3.5-turbo scores higher in 2 categories and Llama 3.1-8B in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Llama 3.1-8B leads 29.7 to 25.3.
- The biggest single-benchmark swing is BigCodeBench Complete: 50.6% for GPT-3.5-turbo and 40.5% for Llama 3.1-8B.
- Llama 3.1-8B is cheaper at $0.05 / $0.08 per million input/output tokens, against $0.50 / $1.50 for GPT-3.5-turbo.
- Llama 3.1-8B accepts more context: 128K tokens versus 16K.
- Llama 3.1-8B has downloadable open weights; the other is API-only.
Side by side
| GPT-3.5-turbo | Llama 3.1-8B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 23.2 | 23.0 |
| Released | 2023-03-01 | 2024-07-23 |
| Weights | Proprietary | Open |
| Context window | 16K | 128K |
| Max output | 4K | 4K |
| Input $ / M tokens | $0.50 | $0.05 |
| Output $ / M tokens | $1.50 | $0.08 |
| Results tracked | 44 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-3.5-turbo leads
GPT-3.5-turbo: 23.9 (#331), Llama 3.1-8B: 20.2 (#340)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| WeirdML | 3.5% | 1.7% |
| BigCodeBench Instruct | 39.1% | 32.8% |
| LMArena Coding | 1136 | 1195 |
| BigCodeBench Complete | 50.6% | 40.5% |
| HumanEval+ | 70.7% | 62.8% |
| MBPP+ | 69.7% | 55.6% |
| SciCode | — | 13.2% |
Agentic & Tool Use Not comparable
GPT-3.5-turbo: —, Llama 3.1-8B: 22.5 (#131)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 25.8% |
| BALROG | — | 15.1% |
| METR Time Horizons | 21.5% | — |
Reasoning Llama 3.1-8B leads
GPT-3.5-turbo: 13.8 (#332), Llama 3.1-8B: 14.9 (#321)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | 1108 | 1175 |
| DTBench | 48.5% | 50.9% |
| LMCA | 9.7% | 5.4% |
| Epoch Capabilities Index | 118.55 | 116.57 |
| CritPt | — | 0% |
| Mystery Game Puzzles | 3% | — |
| Adversarial NLI | 58.1% | — |
| BIG-Bench Hard | 61.6% | — |
| CommonsenseQA 2.0 | 57% | — |
| ForecastBench | 50.4 | — |
| PIQA | — | 81.2% |
| WinoGrande | 81.6% | — |
Math Llama 3.1-8B leads
GPT-3.5-turbo: 6.3 (#327), Llama 3.1-8B: 10.2 (#317)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.2% | 1.7% |
| LMArena Math | 1142 | 1179 |
| MATH Level 5 | 15.9% | 22.9% |
| GSM8K | 57.8% | 82.4% |
| FrontierMath (Tiers 1-3) | 0% | — |
| Omni-MATH | — | 13.7% |
Knowledge GPT-3.5-turbo leads
GPT-3.5-turbo: 10.0 (#303), Llama 3.1-8B: 8.0 (#307)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| GPQA Diamond | 28% | 27% |
| LMArena Expert | 1070 | 1144 |
| BoolQ | 87% | 82.8% |
| MMLU | 71.4% | 56.1% |
| MMLU-Pro | — | 40.6% |
| GPQA (HELM) | — | 24.7% |
| ARC (AI2) Challenge | 87.4% | — |
| OpenBookQA | 86% | — |
| TriviaQA | 85.8% | — |
Multilingual Llama 3.1-8B leads
GPT-3.5-turbo: 31.5 (#258), Llama 3.1-8B: 34.0 (#249)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| LMArena Non-English | 1108 | 1148 |
| LMArena Chinese | 1075 | 1151 |
| LMArena French | 1118 | 1177 |
| LMArena German | 1090 | 1144 |
| LMArena Japanese | 1043 | 1061 |
| LMArena Korean | 1019 | 1053 |
| LMArena Russian | 1123 | 1158 |
| LMArena Spanish | 1121 | 1169 |
Instruction Following Llama 3.1-8B leads
GPT-3.5-turbo: 57.9 (#262), Llama 3.1-8B: 58.9 (#258)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| LMArena Instruction Following | 1119 | 1159 |
| IFEval | — | 74.3% |
Long Context Llama 3.1-8B leads
GPT-3.5-turbo: 34.0 (#254), Llama 3.1-8B: 35.8 (#238)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| LMArena Longer Query | 1121 | 1182 |
Writing & Preference Llama 3.1-8B leads
GPT-3.5-turbo: 25.3 (#305), Llama 3.1-8B: 29.7 (#290)
| Benchmark | GPT-3.5-turbo | Llama 3.1-8B |
|---|---|---|
| LMArena Text | 1125 | 1187 |
| LMArena Creative Writing | 1092 | 1154 |
| EQ-Bench Creative Writing | 451 | 713 |
| LMArena Multi-Turn | 1117 | 1172 |
| WildBench | — | 68.7% |
Frequently asked questions
Is GPT-3.5-turbo better than Llama 3.1-8B?
GPT-3.5-turbo and Llama 3.1-8B score almost the same on the Noometry Index (23.2 vs 23.0), so choose on price, context window or the category you care about most.
Which is cheaper, GPT-3.5-turbo or Llama 3.1-8B?
Llama 3.1-8B is cheaper. It lists at $0.05 per million input tokens and $0.08 per million output tokens; GPT-3.5-turbo lists at $0.50 and $1.50.
Is GPT-3.5-turbo or Llama 3.1-8B better for coding?
GPT-3.5-turbo scores higher on coding benchmarks: 23.9 versus 20.2 in the Noometry coding category.
Which has the bigger context window?
Llama 3.1-8B does, with 128K tokens against 16K.
How many benchmarks do GPT-3.5-turbo and Llama 3.1-8B share?
33 benchmarks have published results for both models. GPT-3.5-turbo has 44 scored results on Noometry and Llama 3.1-8B has 43.