Model comparison
GPT-4o mini vs Llama 3.1-8B
GPT-4o mini is the stronger model overall, scoring 25.5 to 23.0 on the Noometry Index. Llama 3.1-8B costs 4.6× less per token, which makes it the better buy when GPT-4o mini's lead doesn't matter for your workload.
Last verified . 40 shared benchmarks.
MetaSummary
- They share 40 benchmarks with published results for both. GPT-4o mini scores higher in 8 categories and Llama 3.1-8B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-4o mini leads 39.5 to 29.7.
- The biggest single-benchmark swing is MATH Level 5: 52.6% for GPT-4o mini and 22.9% for Llama 3.1-8B.
- Llama 3.1-8B is cheaper at $0.05 / $0.08 per million input/output tokens, against $0.15 / $0.60 for GPT-4o mini.
- Llama 3.1-8B has downloadable open weights; the other is API-only.
Side by side
| GPT-4o mini | Llama 3.1-8B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 25.5 | 23.0 |
| Released | 2024-07-18 | 2024-07-23 |
| Weights | Proprietary | Open |
| Context window | 128K | 128K |
| Max output | 16K | 4K |
| Input $ / M tokens | $0.15 | $0.05 |
| Output $ / M tokens | $0.60 | $0.08 |
| Results tracked | 60 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4o mini leads
GPT-4o mini: 22.0 (#335), Llama 3.1-8B: 20.2 (#340)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| WeirdML | 11.8% | 1.7% |
| BigCodeBench Instruct | 46.1% | 32.8% |
| LMArena Coding | 1290 | 1195 |
| BigCodeBench Complete | 57.4% | 40.5% |
| HumanEval+ | 83.5% | 62.8% |
| MBPP+ | 72.2% | 55.6% |
| Aider Polyglot | 3.6% | — |
| SciCode | — | 13.2% |
| LiveBench Coding | 43.1% | — |
Agentic & Tool Use GPT-4o mini leads
GPT-4o mini: 27.5 (#101), Llama 3.1-8B: 22.5 (#131)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| BALROG | 17.4% | 15.1% |
| Berkeley Function Calling Leaderboard | — | 25.8% |
Reasoning Llama 3.1-8B leads
GPT-4o mini: 8.7 (#347), Llama 3.1-8B: 14.9 (#321)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | 1267 | 1175 |
| DTBench | 54.4% | 50.9% |
| LMCA | 10.4% | 5.4% |
| Epoch Capabilities Index | 126.56 | 116.57 |
| PIQA | 88.7% | 81.2% |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 10.7% | — |
| Kagi LLM Benchmark | 28.8% | — |
| CritPt | — | 0% |
| LiveBench Reasoning | 32.8% | — |
| Mystery Game Puzzles | 12% | — |
| LiveBench Data Analysis | 50% | — |
| LiveBench | 41.3% | — |
Math Too close to call
GPT-4o mini: 10.4 (#314), Llama 3.1-8B: 10.2 (#317)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 6.9% | 1.7% |
| Omni-MATH | 28% | 13.7% |
| LMArena Math | 1267 | 1179 |
| MATH Level 5 | 52.6% | 22.9% |
| GSM8K | 91.3% | 82.4% |
| FrontierMath (Tiers 1-3) | 0.7% | — |
| LiveBench Math | 36.3% | — |
Knowledge GPT-4o mini leads
GPT-4o mini: 17.7 (#284), Llama 3.1-8B: 8.0 (#307)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| GPQA Diamond | 37.7% | 27% |
| MMLU-Pro | 60.3% | 40.6% |
| GPQA (HELM) | 36.8% | 24.7% |
| LMArena Expert | 1235 | 1144 |
| BoolQ | 88.7% | 82.8% |
| MMLU | 81.8% | 56.1% |
| SimpleQA Verified | 8.3% | — |
| Confabulations | 37.2% | — |
Multimodal Not comparable
GPT-4o mini: 25.9 (#122), Llama 3.1-8B: —
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| LMArena Vision | 1066 | — |
| Video-MME | 64.8% | — |
| GeoBench | 64% | — |
| VPCT | 34% | — |
Multilingual GPT-4o mini leads
GPT-4o mini: 42.0 (#199), Llama 3.1-8B: 34.0 (#249)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| LMArena Non-English | 1266 | 1148 |
| LMArena Chinese | 1265 | 1151 |
| LMArena French | 1297 | 1177 |
| LMArena German | 1272 | 1144 |
| LMArena Japanese | 1216 | 1061 |
| LMArena Korean | 1195 | 1053 |
| LMArena Russian | 1275 | 1158 |
| LMArena Spanish | 1276 | 1169 |
Instruction Following GPT-4o mini leads
GPT-4o mini: 61.9 (#239), Llama 3.1-8B: 58.9 (#258)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| IFEval | 78.2% | 74.3% |
| LMArena Instruction Following | 1258 | 1159 |
| LiveBench Instruction Following | 56.8% | — |
Long Context GPT-4o mini leads
GPT-4o mini: 39.1 (#186), Llama 3.1-8B: 35.8 (#238)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| LMArena Longer Query | 1289 | 1182 |
Writing & Preference GPT-4o mini leads
GPT-4o mini: 39.5 (#248), Llama 3.1-8B: 29.7 (#290)
| Benchmark | GPT-4o mini | Llama 3.1-8B |
|---|---|---|
| LMArena Text | 1286 | 1187 |
| LMArena Creative Writing | 1268 | 1154 |
| EQ-Bench Creative Writing | 873 | 713 |
| WildBench | 79.1% | 68.7% |
| LMArena Multi-Turn | 1285 | 1172 |
| Short-Story Creative Writing | 67.2% | — |
| LiveBench Language | 28.6% | — |
Frequently asked questions
Is GPT-4o mini better than Llama 3.1-8B?
GPT-4o mini is the stronger model overall, scoring 25.5 to 23.0 on the Noometry Index. Llama 3.1-8B costs 4.6× less per token, which makes it the better buy when GPT-4o mini's lead doesn't matter for your workload.
Which is cheaper, GPT-4o mini 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-4o mini lists at $0.15 and $0.60.
Is GPT-4o mini or Llama 3.1-8B better for coding?
GPT-4o mini scores higher on coding benchmarks: 22.0 versus 20.2 in the Noometry coding category.
Which has the bigger context window?
Both accept 128K tokens.
How many benchmarks do GPT-4o mini and Llama 3.1-8B share?
40 benchmarks have published results for both models. GPT-4o mini has 60 scored results on Noometry and Llama 3.1-8B has 43.