Model comparison
Llama-3.3-70B-Instruct vs Magistral Medium
Magistral Medium is the stronger model overall, scoring 35.2 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 18× less per token, which makes it the better buy when Magistral Medium's lead doesn't matter for your workload.
Last verified . 19 shared benchmarks.
Meta
Mistral AISummary
- They share 19 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 4 categories and Magistral Medium in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Magistral Medium leads 35.1 to 15.3.
- The biggest single-benchmark swing is SciCode: 26% for Llama-3.3-70B-Instruct and 39.2% for Magistral Medium.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $2 / $5 for Magistral Medium.
- Magistral Medium accepts more context: 262K tokens versus 128K.
Side by side
| Llama-3.3-70B-Instruct | Magistral Medium | |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 30.6 | 35.2 |
| Released | 2024-12-06 | 2025-03-17 |
| Weights | Open | Open |
| Context window | 128K | 262K |
| Max output | 4K | 16K |
| Input $ / M tokens | $0.10 | $2 |
| Output $ / M tokens | $0.32 | $5 |
| Results tracked | 43 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Magistral Medium leads
Llama-3.3-70B-Instruct: 31.0 (#290), Magistral Medium: 39.1 (#161)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| SciCode | 26% | 39.2% |
| LMArena Coding | 1268 | 1319 |
| WeirdML | 14.4% | — |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
| BigCodeBench Complete | 57.5% | — |
Agentic & Tool Use Not comparable
Llama-3.3-70B-Instruct: 25.8 (#105), Magistral Medium: —
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | — |
| BALROG | 23% | — |
Reasoning Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 14.1 (#327), Magistral Medium: 8.6 (#348)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| CritPt | 0% | 0.3% |
| LMArena Hard Prompts | 1257 | 1267 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | 19.9% | — |
| Kagi LLM Benchmark | — | 16.2% |
| ARC-AGI-1 | — | 6.1% |
| LiveBench Reasoning | 50.8% | — |
| DTBench | 59.5% | — |
| LiveBench Data Analysis | 49.5% | — |
| LMCA | 17.5% | — |
| Epoch Capabilities Index | 127.33 | — |
| ForecastBench | 58.6 | — |
| LiveBench | 50.2% | — |
Math Magistral Medium leads
Llama-3.3-70B-Instruct: 15.3 (#298), Magistral Medium: 35.1 (#189)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| LMArena Math | 1267 | 1250 |
| OTIS Mock AIME 2024-2025 | 5.1% | — |
| LiveBench Math | 42.2% | — |
| MATH Level 5 | 41.6% | — |
Knowledge Magistral Medium leads
Llama-3.3-70B-Instruct: 30.6 (#226), Magistral Medium: 33.5 (#202)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| LMArena Expert | 1225 | 1223 |
| GPQA Diamond | 47.4% | — |
| Confabulations | 22.8% | — |
| Vectara Hallucination Rate | 4.1% | — |
| MMLU | 86.3% | — |
Multilingual Too close to call
Llama-3.3-70B-Instruct: 39.9 (#220), Magistral Medium: 39.6 (#224)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| LMArena Non-English | 1236 | 1232 |
| LMArena Chinese | 1217 | 1227 |
| LMArena French | 1281 | 1267 |
| LMArena German | 1251 | 1248 |
| LMArena Japanese | 1150 | 1175 |
| LMArena Korean | 1143 | 1125 |
| LMArena Russian | 1252 | 1224 |
| LMArena Spanish | 1270 | 1271 |
Instruction Following Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 71.1 (#157), Magistral Medium: 66.0 (#211)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| LMArena Instruction Following | 1242 | 1254 |
| LiveBench Instruction Following | 82.7% | — |
Long Context Magistral Medium leads
Llama-3.3-70B-Instruct: 26.4 (#295), Magistral Medium: 39.3 (#183)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| LMArena Longer Query | 1256 | 1295 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Llama-3.3-70B-Instruct leads
Llama-3.3-70B-Instruct: 47.6 (#207), Magistral Medium: 46.3 (#219)
| Benchmark | Llama-3.3-70B-Instruct | Magistral Medium |
|---|---|---|
| LMArena Text | 1274 | 1255 |
| LMArena Creative Writing | 1250 | 1245 |
| LMArena Multi-Turn | 1280 | 1275 |
| LiveBench Language | 39.2% | — |
Frequently asked questions
Is Llama-3.3-70B-Instruct better than Magistral Medium?
Magistral Medium is the stronger model overall, scoring 35.2 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 18× less per token, which makes it the better buy when Magistral Medium's lead doesn't matter for your workload.
Which is cheaper, Llama-3.3-70B-Instruct or Magistral Medium?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; Magistral Medium lists at $2 and $5.
Is Llama-3.3-70B-Instruct or Magistral Medium better for coding?
Magistral Medium scores higher on coding benchmarks: 39.1 versus 31.0 in the Noometry coding category.
Which has the bigger context window?
Magistral Medium does, with 262K tokens against 128K.
How many benchmarks do Llama-3.3-70B-Instruct and Magistral Medium share?
19 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and Magistral Medium has 22.