Model comparison
Llama 3.2 1B vs MiniMax-M3
MiniMax-M3 is the stronger model overall, scoring 43.5 to 19.9 on the Noometry Index. Llama 3.2 1B costs 7.4× less per token, which makes it the better buy when MiniMax-M3's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
Meta
MiniMaxSummary
- They share 17 benchmarks with published results for both. Llama 3.2 1B scores higher in 0 categories and MiniMax-M3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 7.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 0.6% for Llama 3.2 1B and 71.1% for MiniMax-M3.
- Llama 3.2 1B is cheaper at $0.027 / $0.20 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M3.
- MiniMax-M3 accepts more context: 1M tokens versus 60K.
Side by side
| Llama 3.2 1B | MiniMax-M3 | |
|---|---|---|
| Provider | Meta | MiniMax |
| Noometry Index | 19.9 | 43.5 |
| Released | 2024-09-24 | 2026-06-01 |
| Weights | Open | Open |
| Context window | 60K | 1M |
| Max output | 54K | 512K |
| Input $ / M tokens | $0.027 | $0.30 |
| Output $ / M tokens | $0.20 | $1.20 |
| Results tracked | 22 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M3 leads
Llama 3.2 1B: 20.6 (#333), MiniMax-M3: 41.5 (#111)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| LMArena Coding | 1071 | 1470 |
| FrontierCode | — | 14.7% |
| LMArena WebDev | — | 1482 |
| SciCode | — | 47.1% |
| BigCodeBench Instruct | 8.2% | — |
| BigCodeBench Complete | 11.3% | — |
| ALE-Bench | — | 640.02 |
Agentic & Tool Use MiniMax-M3 leads
Llama 3.2 1B: 14.5 (#142), MiniMax-M3: 22.5 (#121)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| APEX-Agents | — | 37.7% |
| Berkeley Function Calling Leaderboard | 10.8% | — |
| OSWorld 2.0 | — | 4.6% |
| BALROG | 6.6% | — |
| GBAEval | — | 0.9% |
| Vending-Bench 2 | — | 2,158 |
Reasoning MiniMax-M3 leads
Llama 3.2 1B: 15.6 (#304), MiniMax-M3: 29.3 (#83)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| Chess Puzzles | 0% | 14% |
| LMArena Hard Prompts | 1044 | 1447 |
| Epoch Capabilities Index | 101.95 | 147 |
| SimpleBench | — | 45.8% |
| NYT Connections (extended) | — | 65.1% |
| CritPt | — | 3.7% |
| Mystery Game Puzzles | — | 8% |
| DTBench | — | 78.9% |
| LMCA | — | 33.7% |
| Surface Evolver Bench | — | 55% |
| ForecastBench | — | 61.4 |
Math MiniMax-M3 leads
Llama 3.2 1B: 9.7 (#308), MiniMax-M3: 39.1 (#83)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.6% | 71.1% |
| LMArena Math | 1086 | 1432 |
| ProofBench | — | 18% |
Knowledge MiniMax-M3 leads
Llama 3.2 1B: 7.2 (#307), MiniMax-M3: 58.4 (#32)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 23.9% | 90.9% |
| LMArena Expert | 1007 | 1462 |
Multimodal Not comparable
Llama 3.2 1B: —, MiniMax-M3: 39.5 (#45)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| LMArena Vision | — | 1254 |
| LMArena Document | — | 1435 |
Multilingual MiniMax-M3 leads
Llama 3.2 1B: 23.9 (#287), MiniMax-M3: 53.2 (#70)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 973 | 1420 |
| LMArena Chinese | 958 | 1465 |
| LMArena German | 1014 | 1424 |
| LMArena Russian | 941 | 1430 |
| LMArena French | — | 1445 |
| LMArena Japanese | — | 1381 |
| LMArena Korean | — | 1372 |
| LMArena Spanish | — | 1431 |
Instruction Following MiniMax-M3 leads
Llama 3.2 1B: 52.6 (#285), MiniMax-M3: 75.7 (#56)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1031 | 1433 |
Long Context MiniMax-M3 leads
Llama 3.2 1B: 31.9 (#269), MiniMax-M3: 44.2 (#68)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1050 | 1445 |
Writing & Preference MiniMax-M3 leads
Llama 3.2 1B: 21.4 (#305), MiniMax-M3: 62.3 (#79)
| Benchmark | Llama 3.2 1B | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1055 | 1433 |
| LMArena Creative Writing | 1033 | 1405 |
| LMArena Multi-Turn | 1030 | 1441 |
| EQ-Bench Creative Writing | 200 | — |
| EQ-Bench 4 | — | 1150 |
Frequently asked questions
Is Llama 3.2 1B better than MiniMax-M3?
MiniMax-M3 is the stronger model overall, scoring 43.5 to 19.9 on the Noometry Index. Llama 3.2 1B costs 7.4× less per token, which makes it the better buy when MiniMax-M3's lead doesn't matter for your workload.
Which is cheaper, Llama 3.2 1B or MiniMax-M3?
Llama 3.2 1B is cheaper. It lists at $0.027 per million input tokens and $0.20 per million output tokens; MiniMax-M3 lists at $0.30 and $1.20.
Is Llama 3.2 1B or MiniMax-M3 better for coding?
MiniMax-M3 scores higher on coding benchmarks: 41.5 versus 20.6 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 60K.
How many benchmarks do Llama 3.2 1B and MiniMax-M3 share?
17 benchmarks have published results for both models. Llama 3.2 1B has 22 scored results on Noometry and MiniMax-M3 has 41.