Model comparison
Llama 3.1-70B vs MiniMax-M3
MiniMax-M3 is the stronger model overall, scoring 43.8 to 29.6 on the Noometry Index.
Last verified . 22 shared benchmarks.
Meta
MiniMaxSummary
- They share 22 benchmarks with published results for both. Llama 3.1-70B scores higher in 1 category and MiniMax-M3 in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 24.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 3.6% for Llama 3.1-70B and 71.1% for MiniMax-M3.
- Llama 3.1-70B is cheaper at $0.40 / $0.40 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M3.
- MiniMax-M3 accepts more context: 1M tokens versus 128K.
Side by side
| Llama 3.1-70B | MiniMax-M3 | |
|---|---|---|
| Provider | Meta | MiniMax |
| Noometry Index | 29.6 | 43.8 |
| Released | 2024-07-23 | 2026-06-01 |
| Weights | Open | Open |
| Context window | 128K | 1M |
| Max output | 4K | 512K |
| Input $ / M tokens | $0.40 | $0.30 |
| Output $ / M tokens | $0.40 | $1.20 |
| Results tracked | 35 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M3 leads
Llama 3.1-70B: 30.3 (#296), MiniMax-M3: 41.8 (#118)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| LMArena Coding | 1260 | 1469 |
| FrontierCode | — | 14.7% |
| LMArena WebDev | — | 1482 |
| SciCode | — | 47.1% |
| WeirdML | 9% | — |
| BigCodeBench Instruct | 46.1% | — |
| BigCodeBench Complete | 54.8% | — |
| ALE-Bench | — | 640.02 |
Agentic & Tool Use Llama 3.1-70B leads
Llama 3.1-70B: 25.1 (#112), MiniMax-M3: 22.6 (#130)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| APEX-Agents | — | 37.7% |
| OSWorld 2.0 | — | 4.6% |
| TheAgentCompany | 6.9% | — |
| BALROG | 27.9% | — |
| GBAEval | — | 0.9% |
| Vending-Bench 2 | — | 2,158 |
Reasoning MiniMax-M3 leads
Llama 3.1-70B: 21.6 (#220), MiniMax-M3: 30.1 (#87)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1447 |
| DTBench | 60% | 78.9% |
| LMCA | 14.8% | 33.7% |
| Epoch Capabilities Index | 125.92 | 146.95 |
| SimpleBench | — | 45.8% |
| NYT Connections (extended) | — | 65.1% |
| CritPt | — | 3.7% |
| Chess Puzzles | — | 14% |
| Mystery Game Puzzles | — | 8% |
| Surface Evolver Bench | — | 55% |
| ForecastBench | — | 61.4 |
Math MiniMax-M3 leads
Llama 3.1-70B: 13.5 (#304), MiniMax-M3: 40.0 (#95)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 3.6% | 71.1% |
| LMArena Math | 1252 | 1429 |
| ProofBench | — | 18% |
| Omni-MATH | 21% | — |
| MATH Level 5 | 36.7% | — |
Knowledge MiniMax-M3 leads
Llama 3.1-70B: 24.2 (#269), MiniMax-M3: 58.4 (#35)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 44.2% | 90.9% |
| LMArena Expert | 1209 | 1461 |
| MMLU-Pro | 65.3% | — |
| GPQA (HELM) | 42.6% | — |
| MMLU | 80.1% | — |
Multimodal Not comparable
Llama 3.1-70B: —, MiniMax-M3: 40.2 (#51)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| LMArena Vision | — | 1253 |
| LMArena Document | — | 1435 |
Multilingual MiniMax-M3 leads
Llama 3.1-70B: 38.8 (#225), MiniMax-M3: 53.0 (#75)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1219 | 1420 |
| LMArena Chinese | 1215 | 1463 |
| LMArena French | 1261 | 1447 |
| LMArena German | 1222 | 1426 |
| LMArena Japanese | 1132 | 1381 |
| LMArena Korean | 1140 | 1372 |
| LMArena Russian | 1234 | 1428 |
| LMArena Spanish | 1253 | 1432 |
Instruction Following MiniMax-M3 leads
Llama 3.1-70B: 65.3 (#223), MiniMax-M3: 75.5 (#62)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1231 | 1433 |
| IFEval | 82.1% | — |
Long Context MiniMax-M3 leads
Llama 3.1-70B: 37.6 (#214), MiniMax-M3: 44.2 (#72)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1241 | 1445 |
Writing & Preference MiniMax-M3 leads
Llama 3.1-70B: 35.4 (#267), MiniMax-M3: 62.1 (#83)
| Benchmark | Llama 3.1-70B | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1261 | 1433 |
| LMArena Creative Writing | 1232 | 1404 |
| LMArena Multi-Turn | 1256 | 1442 |
| EQ-Bench Creative Writing | 784 | — |
| WildBench | 75.8% | — |
| EQ-Bench 4 | — | 1150 |
Frequently asked questions
Is Llama 3.1-70B better than MiniMax-M3?
MiniMax-M3 is the stronger model overall, scoring 43.8 to 29.6 on the Noometry Index.
Which is cheaper, Llama 3.1-70B or MiniMax-M3?
Llama 3.1-70B is cheaper. It lists at $0.40 per million input tokens and $0.40 per million output tokens; MiniMax-M3 lists at $0.30 and $1.20.
Is Llama 3.1-70B or MiniMax-M3 better for coding?
MiniMax-M3 scores higher on coding benchmarks: 41.8 versus 30.3 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 128K.
How many benchmarks do Llama 3.1-70B and MiniMax-M3 share?
22 benchmarks have published results for both models. Llama 3.1-70B has 35 scored results on Noometry and MiniMax-M3 has 41.