Model comparison
Llama-3.3-70B-Instruct vs Ministral 3B
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.3 to 25.6 on the Noometry Index. Ministral 3B costs 3.9× less per token, which makes it the better buy when Llama-3.3-70B-Instruct's lead doesn't matter for your workload.
Last verified . 6 shared benchmarks.
Meta
Mistral AISummary
- They share 6 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 1 category and Ministral 3B in 2 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama-3.3-70B-Instruct leads 30.5 to 10.4.
- The biggest single-benchmark swing is MATH Level 5: 41.6% for Llama-3.3-70B-Instruct and 14.4% for Ministral 3B.
- Ministral 3B is cheaper at $0.04 / $0.04 per million input/output tokens, against $0.10 / $0.32 for Llama-3.3-70B-Instruct.
Side by side
| Llama-3.3-70B-Instruct | Ministral 3B | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 30.3 | 25.6 |
| Released | 2024-12-06 | 2024-10-01 |
| Weights | Open | Open |
| Context window | 128K | 128K |
| Max output | 4K | 128K |
| Input $ / M tokens | $0.10 | $0.04 |
| Output $ / M tokens | $0.32 | $0.04 |
| Results tracked | 43 | 6 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Llama-3.3-70B-Instruct: 30.6 (#284), Ministral 3B: —
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| SciCode | 26% | — |
| WeirdML | 14.4% | — |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
| LMArena Coding | 1268 | — |
| BigCodeBench Complete | 57.5% | — |
Agentic & Tool Use Not comparable
Llama-3.3-70B-Instruct: 25.6 (#100), Ministral 3B: —
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | — |
| BALROG | 23% | — |
Reasoning Ministral 3B leads
Llama-3.3-70B-Instruct: 13.7 (#325), Ministral 3B: 17.7 (#278)
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| DTBench | 59.5% | 51.7% |
| LMCA | 17.5% | 5.5% |
| Epoch Capabilities Index | 127.32 | 118.08 |
| SimpleBench | 19.9% | — |
| CritPt | 0% | — |
| LiveBench Reasoning | 50.8% | — |
| LMArena Hard Prompts | 1257 | — |
| LiveBench Data Analysis | 49.5% | — |
| ForecastBench | 58.6 | — |
| LiveBench | 50.2% | — |
Math Ministral 3B leads
Llama-3.3-70B-Instruct: 14.3 (#292), Ministral 3B: 24.5 (#257)
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| MATH Level 5 | 41.6% | 14.4% |
| OTIS Mock AIME 2024-2025 | 5.1% | — |
| LiveBench Math | 42.2% | — |
| LMArena Math | 1268 | — |
Knowledge Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 30.5 (#220), Ministral 3B: 10.4 (#297)
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| GPQA Diamond | 47.4% | 25.3% |
| Vectara Hallucination Rate | 4.1% | 7.3% |
| Confabulations | 22.8% | — |
| LMArena Expert | 1225 | — |
| MMLU | 86.3% | — |
Multilingual Not comparable
Llama-3.3-70B-Instruct: 40.1 (#215), Ministral 3B: —
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| LMArena Non-English | 1236 | — |
| LMArena Chinese | 1217 | — |
| LMArena French | 1280 | — |
| LMArena German | 1251 | — |
| LMArena Japanese | 1149 | — |
| LMArena Korean | 1142 | — |
| LMArena Russian | 1252 | — |
| LMArena Spanish | 1269 | — |
Instruction Following Not comparable
Llama-3.3-70B-Instruct: 71.3 (#152), Ministral 3B: —
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| LiveBench Instruction Following | 82.7% | — |
| LMArena Instruction Following | 1242 | — |
Long Context Not comparable
Llama-3.3-70B-Instruct: 26.4 (#290), Ministral 3B: —
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| Fiction.LiveBench | 33.3% | — |
| LMArena Longer Query | 1256 | — |
Writing & Preference Not comparable
Llama-3.3-70B-Instruct: 47.8 (#203), Ministral 3B: —
| Benchmark | Llama-3.3-70B-Instruct | Ministral 3B |
|---|---|---|
| LMArena Text | 1274 | — |
| LMArena Creative Writing | 1250 | — |
| LMArena Multi-Turn | 1279 | — |
| LiveBench Language | 39.2% | — |
Frequently asked questions
Is Llama-3.3-70B-Instruct better than Ministral 3B?
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.3 to 25.6 on the Noometry Index. Ministral 3B costs 3.9× less per token, which makes it the better buy when Llama-3.3-70B-Instruct's lead doesn't matter for your workload.
Which is cheaper, Llama-3.3-70B-Instruct or Ministral 3B?
Ministral 3B is cheaper. It lists at $0.04 per million input tokens and $0.04 per million output tokens; Llama-3.3-70B-Instruct lists at $0.10 and $0.32.
Which has the bigger context window?
Both accept 128K tokens.
How many benchmarks do Llama-3.3-70B-Instruct and Ministral 3B share?
6 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and Ministral 3B has 6.