Model comparison
Llama 3.1-70B vs Mistral Large
Mistral Large is the stronger model overall, scoring 31.9 to 29.6 on the Noometry Index. Llama 3.1-70B costs 7.5× less per token, which makes it the better buy when Mistral Large's lead doesn't matter for your workload.
Last verified . 32 shared benchmarks.
Meta
Mistral AISummary
- They share 32 benchmarks with published results for both. Llama 3.1-70B scores higher in 1 category and Mistral Large in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Mistral Large leads 30.1 to 24.2.
- The biggest single-benchmark swing is BigCodeBench Complete: 54.8% for Llama 3.1-70B and 38.3% for Mistral Large.
- Llama 3.1-70B is cheaper at $0.40 / $0.40 per million input/output tokens, against $2 / $6 for Mistral Large.
- Mistral Large accepts more context: 131K tokens versus 128K.
Side by side
| Llama 3.1-70B | Mistral Large | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 29.6 | 31.9 |
| Released | 2024-07-23 | 2024-02-26 |
| Weights | Open | Open |
| Context window | 128K | 131K |
| Max output | 4K | 16K |
| Input $ / M tokens | $0.40 | $2 |
| Output $ / M tokens | $0.40 | $6 |
| Results tracked | 35 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Large leads
Llama 3.1-70B: 30.3 (#296), Mistral Large: 34.3 (#240)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| BigCodeBench Instruct | 46.1% | 30% |
| LMArena Coding | 1260 | 1277 |
| BigCodeBench Complete | 54.8% | 38.3% |
| SciCode | — | 36.2% |
| WeirdML | 9% | — |
| LiveBench Coding | — | 47.1% |
| ALE-Bench | — | 264.7 |
| HumanEval+ | — | 62.2% |
| MBPP+ | — | 59.5% |
Agentic & Tool Use Mistral Large leads
Llama 3.1-70B: 25.1 (#112), Mistral Large: 28.6 (#89)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 38.4% |
| TheAgentCompany | 6.9% | — |
| BALROG | 27.9% | — |
Reasoning Llama 3.1-70B leads
Llama 3.1-70B: 21.6 (#220), Mistral Large: 15.8 (#310)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1257 |
| DTBench | 60% | 65.1% |
| LMCA | 14.8% | 16.7% |
| Epoch Capabilities Index | 125.92 | 128.52 |
| SimpleBench | — | 22.5% |
| CritPt | — | 0% |
| LiveBench Reasoning | — | 43.5% |
| LiveBench Data Analysis | — | 50.1% |
| ForecastBench | — | 57.1 |
| LiveBench | — | 48.4% |
Math Mistral Large leads
Llama 3.1-70B: 13.5 (#304), Mistral Large: 18.2 (#291)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 3.6% | 8.5% |
| Omni-MATH | 21% | 28.1% |
| LMArena Math | 1252 | 1262 |
| MATH Level 5 | 36.7% | 50.3% |
| LiveBench Math | — | 42.5% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge Mistral Large leads
Llama 3.1-70B: 24.2 (#269), Mistral Large: 30.1 (#230)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| GPQA Diamond | 44.2% | 51.3% |
| MMLU-Pro | 65.3% | 59.9% |
| GPQA (HELM) | 42.6% | 43.5% |
| LMArena Expert | 1209 | 1232 |
| MMLU | 80.1% | 80% |
| Confabulations | — | 21.4% |
| Vectara Hallucination Rate | — | 4.5% |
Multilingual Mistral Large leads
Llama 3.1-70B: 38.8 (#225), Mistral Large: 40.0 (#219)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| LMArena Non-English | 1219 | 1237 |
| LMArena Chinese | 1215 | 1240 |
| LMArena French | 1261 | 1325 |
| LMArena German | 1222 | 1254 |
| LMArena Japanese | 1132 | 1188 |
| LMArena Korean | 1140 | 1202 |
| LMArena Russian | 1234 | 1257 |
| LMArena Spanish | 1253 | 1268 |
Instruction Following Mistral Large leads
Llama 3.1-70B: 65.3 (#223), Mistral Large: 67.9 (#191)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| IFEval | 82.1% | 87.7% |
| LMArena Instruction Following | 1231 | 1249 |
| LiveBench Instruction Following | — | 67.9% |
Long Context Too close to call
Llama 3.1-70B: 37.6 (#214), Mistral Large: 38.3 (#199)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| LMArena Longer Query | 1241 | 1261 |
Writing & Preference Mistral Large leads
Llama 3.1-70B: 35.4 (#267), Mistral Large: 40.7 (#242)
| Benchmark | Llama 3.1-70B | Mistral Large |
|---|---|---|
| LMArena Text | 1261 | 1266 |
| LMArena Creative Writing | 1232 | 1243 |
| EQ-Bench Creative Writing | 784 | 985 |
| WildBench | 75.8% | 80.1% |
| LMArena Multi-Turn | 1256 | 1260 |
| Short-Story Creative Writing | — | 69% |
| LiveBench Language | — | 39.4% |
Frequently asked questions
Is Llama 3.1-70B better than Mistral Large?
Mistral Large is the stronger model overall, scoring 31.9 to 29.6 on the Noometry Index. Llama 3.1-70B costs 7.5× less per token, which makes it the better buy when Mistral Large's lead doesn't matter for your workload.
Which is cheaper, Llama 3.1-70B or Mistral Large?
Llama 3.1-70B is cheaper. It lists at $0.40 per million input tokens and $0.40 per million output tokens; Mistral Large lists at $2 and $6.
Is Llama 3.1-70B or Mistral Large better for coding?
Mistral Large scores higher on coding benchmarks: 34.3 versus 30.3 in the Noometry coding category.
Which has the bigger context window?
Mistral Large does, with 131K tokens against 128K.
How many benchmarks do Llama 3.1-70B and Mistral Large share?
32 benchmarks have published results for both models. Llama 3.1-70B has 35 scored results on Noometry and Mistral Large has 51.