Model comparison
Llama 3.1-70B vs Mistral Small
Mistral Small is the stronger model overall, scoring 33.4 to 29.6 on the Noometry Index.
Last verified . 25 shared benchmarks.
Meta
Mistral AISummary
- They share 25 benchmarks with published results for both. Llama 3.1-70B scores higher in 1 category and Mistral Small in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mistral Small leads 52.5 to 35.4.
- The biggest single-benchmark swing is DTBench: 60% for Llama 3.1-70B and 70.9% for Mistral Small.
- Mistral Small is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.40 / $0.40 for Llama 3.1-70B.
- Mistral Small accepts more context: 262K tokens versus 128K.
Side by side
| Llama 3.1-70B | Mistral Small | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 29.6 | 33.4 |
| Released | 2024-07-23 | 2024-02-26 |
| Weights | Open | Open |
| Context window | 128K | 262K |
| Max output | 4K | 256K |
| Input $ / M tokens | $0.40 | $0.15 |
| Output $ / M tokens | $0.40 | $0.60 |
| Results tracked | 35 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mistral Small leads
Llama 3.1-70B: 30.3 (#296), Mistral Small: 34.0 (#247)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| BigCodeBench Instruct | 46.1% | 36.1% |
| LMArena Coding | 1260 | 1362 |
| BigCodeBench Complete | 54.8% | 46.6% |
| SciCode | — | 26.5% |
| WeirdML | 9% | — |
| LiveBench Coding | — | 36.2% |
| ALE-Bench | — | 497.62 |
Agentic & Tool Use Mistral Small leads
Llama 3.1-70B: 25.1 (#112), Mistral Small: 28.1 (#93)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 37.1% |
| TheAgentCompany | 6.9% | — |
| BALROG | 27.9% | — |
Reasoning Llama 3.1-70B leads
Llama 3.1-70B: 21.6 (#220), Mistral Small: 19.8 (#250)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1335 |
| DTBench | 60% | 70.9% |
| LMCA | 14.8% | 20.6% |
| Kagi LLM Benchmark | — | 37.8% |
| CritPt | — | 0% |
| LiveBench Reasoning | — | 44.8% |
| LiveBench Data Analysis | — | 53.7% |
| Epoch Capabilities Index | 125.92 | — |
| LiveBench | — | 44% |
Math Mistral Small leads
Llama 3.1-70B: 13.5 (#304), Mistral Small: 16.4 (#293)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 3.6% | 5.8% |
| LMArena Math | 1252 | 1341 |
| MATH Level 5 | 36.7% | 46.8% |
| Omni-MATH | 21% | — |
| LiveBench Math | — | 39.9% |
Knowledge Mistral Small leads
Llama 3.1-70B: 24.2 (#269), Mistral Small: 31.0 (#222)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| GPQA Diamond | 44.2% | 47.5% |
| LMArena Expert | 1209 | 1291 |
| MMLU | 80.1% | 68.7% |
| MMLU-Pro | 65.3% | — |
| Vectara Hallucination Rate | — | 5.1% |
| GPQA (HELM) | 42.6% | — |
Multimodal Not comparable
Llama 3.1-70B: —, Mistral Small: 33.5 (#96)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| LMArena Vision | — | 1142 |
Multilingual Mistral Small leads
Llama 3.1-70B: 38.8 (#225), Mistral Small: 45.5 (#169)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| LMArena Non-English | 1219 | 1315 |
| LMArena Chinese | 1215 | 1340 |
| LMArena French | 1261 | 1337 |
| LMArena German | 1222 | 1340 |
| LMArena Japanese | 1132 | 1275 |
| LMArena Korean | 1140 | 1259 |
| LMArena Russian | 1234 | 1324 |
| LMArena Spanish | 1253 | 1346 |
Instruction Following Mistral Small leads
Llama 3.1-70B: 65.3 (#223), Mistral Small: 66.4 (#209)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| LMArena Instruction Following | 1231 | 1310 |
| LiveBench Instruction Following | — | 63.7% |
| IFEval | 82.1% | — |
Long Context Mistral Small leads
Llama 3.1-70B: 37.6 (#214), Mistral Small: 40.4 (#156)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| LMArena Longer Query | 1241 | 1327 |
Writing & Preference Mistral Small leads
Llama 3.1-70B: 35.4 (#267), Mistral Small: 52.5 (#171)
| Benchmark | Llama 3.1-70B | Mistral Small |
|---|---|---|
| LMArena Text | 1261 | 1338 |
| LMArena Creative Writing | 1232 | 1305 |
| LMArena Multi-Turn | 1256 | 1344 |
| EQ-Bench Creative Writing | 784 | — |
| WildBench | 75.8% | — |
| LiveBench Language | — | 30.5% |
Frequently asked questions
Is Llama 3.1-70B better than Mistral Small?
Mistral Small is the stronger model overall, scoring 33.4 to 29.6 on the Noometry Index.
Which is cheaper, Llama 3.1-70B or Mistral Small?
Mistral Small is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Llama 3.1-70B lists at $0.40 and $0.40.
Is Llama 3.1-70B or Mistral Small better for coding?
Mistral Small scores higher on coding benchmarks: 34.0 versus 30.3 in the Noometry coding category.
Which has the bigger context window?
Mistral Small does, with 262K tokens against 128K.
How many benchmarks do Llama 3.1-70B and Mistral Small share?
25 benchmarks have published results for both models. Llama 3.1-70B has 35 scored results on Noometry and Mistral Small has 39.