Model comparison
Llama 3.1-8B vs Ministral 3B
Ministral 3B is the stronger model overall, scoring 26.2 to 23.0 on the Noometry Index. Llama 3.1-8B costs 1.7× less per token, which makes it the better buy when Ministral 3B's lead doesn't matter for your workload.
Last verified . 5 shared benchmarks.
Meta
Mistral AISummary
- They share 5 benchmarks with published results for both. Llama 3.1-8B scores higher in 0 categories and Ministral 3B in 3 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in math, where Ministral 3B leads 26.6 to 10.2.
- The biggest single-benchmark swing is MATH Level 5: 22.9% for Llama 3.1-8B and 14.4% for Ministral 3B.
- Llama 3.1-8B is cheaper at $0.05 / $0.08 per million input/output tokens, against $0.10 / $0.10 for Ministral 3B.
- Ministral 3B accepts more context: 131K tokens versus 128K.
Side by side
| Llama 3.1-8B | Ministral 3B | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 23.0 | 26.2 |
| Released | 2024-07-23 | 2024-10-01 |
| Weights | Open | Open |
| Context window | 128K | 131K |
| Max output | 4K | 262K |
| Input $ / M tokens | $0.05 | $0.10 |
| Output $ / M tokens | $0.08 | $0.10 |
| Results tracked | 43 | 6 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Llama 3.1-8B: 20.2 (#340), Ministral 3B: —
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| SciCode | 13.2% | — |
| WeirdML | 1.7% | — |
| BigCodeBench Instruct | 32.8% | — |
| LMArena Coding | 1195 | — |
| BigCodeBench Complete | 40.5% | — |
| HumanEval+ | 62.8% | — |
| MBPP+ | 55.6% | — |
Agentic & Tool Use Not comparable
Llama 3.1-8B: 22.5 (#131), Ministral 3B: —
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 25.8% | — |
| BALROG | 15.1% | — |
Reasoning Ministral 3B leads
Llama 3.1-8B: 14.9 (#321), Ministral 3B: 18.4 (#282)
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| DTBench | 50.9% | 51.7% |
| LMCA | 5.4% | 5.5% |
| Epoch Capabilities Index | 116.57 | 118.1 |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| LMArena Hard Prompts | 1175 | — |
| PIQA | 81.2% | — |
Math Ministral 3B leads
Llama 3.1-8B: 10.2 (#317), Ministral 3B: 26.6 (#258)
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| MATH Level 5 | 22.9% | 14.4% |
| OTIS Mock AIME 2024-2025 | 1.7% | — |
| Omni-MATH | 13.7% | — |
| LMArena Math | 1179 | — |
| GSM8K | 82.4% | — |
Knowledge Ministral 3B leads
Llama 3.1-8B: 8.0 (#307), Ministral 3B: 10.4 (#302)
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| GPQA Diamond | 27% | 25.3% |
| MMLU-Pro | 40.6% | — |
| Vectara Hallucination Rate | — | 7.3% |
| GPQA (HELM) | 24.7% | — |
| LMArena Expert | 1144 | — |
| BoolQ | 82.8% | — |
| MMLU | 56.1% | — |
Multilingual Not comparable
Llama 3.1-8B: 34.0 (#249), Ministral 3B: —
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| LMArena Non-English | 1148 | — |
| LMArena Chinese | 1151 | — |
| LMArena French | 1177 | — |
| LMArena German | 1144 | — |
| LMArena Japanese | 1061 | — |
| LMArena Korean | 1053 | — |
| LMArena Russian | 1158 | — |
| LMArena Spanish | 1169 | — |
Instruction Following Not comparable
Llama 3.1-8B: 58.9 (#258), Ministral 3B: —
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| IFEval | 74.3% | — |
| LMArena Instruction Following | 1159 | — |
Long Context Not comparable
Llama 3.1-8B: 35.8 (#238), Ministral 3B: —
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| LMArena Longer Query | 1182 | — |
Writing & Preference Not comparable
Llama 3.1-8B: 29.7 (#290), Ministral 3B: —
| Benchmark | Llama 3.1-8B | Ministral 3B |
|---|---|---|
| LMArena Text | 1187 | — |
| LMArena Creative Writing | 1154 | — |
| EQ-Bench Creative Writing | 713 | — |
| WildBench | 68.7% | — |
| LMArena Multi-Turn | 1172 | — |
Frequently asked questions
Is Llama 3.1-8B better than Ministral 3B?
Ministral 3B is the stronger model overall, scoring 26.2 to 23.0 on the Noometry Index. Llama 3.1-8B costs 1.7× less per token, which makes it the better buy when Ministral 3B's lead doesn't matter for your workload.
Which is cheaper, Llama 3.1-8B or Ministral 3B?
Llama 3.1-8B is cheaper. It lists at $0.05 per million input tokens and $0.08 per million output tokens; Ministral 3B lists at $0.10 and $0.10.
Which has the bigger context window?
Ministral 3B does, with 131K tokens against 128K.
How many benchmarks do Llama 3.1-8B and Ministral 3B share?
5 benchmarks have published results for both models. Llama 3.1-8B has 43 scored results on Noometry and Ministral 3B has 6.