Model comparison
Llama 3.1-8B vs MiniMax-M2.7
MiniMax-M2.7 is the stronger model overall, scoring 37.4 to 22.8 on the Noometry Index. Llama 3.1-8B costs 9.1× less per token, which makes it the better buy when MiniMax-M2.7's lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
Meta
MiniMaxSummary
- They share 21 benchmarks with published results for both. Llama 3.1-8B scores higher in 0 categories and MiniMax-M2.7 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where MiniMax-M2.7 leads 59.1 to 29.9.
- The biggest single-benchmark swing is WeirdML: 1.7% for Llama 3.1-8B and 37% for MiniMax-M2.7.
- Llama 3.1-8B is cheaper at $0.05 / $0.08 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.7.
- MiniMax-M2.7 accepts more context: 205K tokens versus 128K.
Side by side
| Llama 3.1-8B | MiniMax-M2.7 | |
|---|---|---|
| Provider | Meta | MiniMax |
| Noometry Index | 22.8 | 37.4 |
| Released | 2024-07-23 | 2026-03-18 |
| Weights | Open | Open |
| Context window | 128K | 205K |
| Max output | 4K | 131K |
| Input $ / M tokens | $0.05 | $0.30 |
| Output $ / M tokens | $0.08 | $1.20 |
| Results tracked | 43 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.7 leads
Llama 3.1-8B: 19.9 (#335), MiniMax-M2.7: 41.4 (#115)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| SciCode | 13.2% | 47% |
| WeirdML | 1.7% | 37% |
| LMArena Coding | 1195 | 1454 |
| LMArena WebDev | — | 1397 |
| BigCodeBench Instruct | 32.8% | — |
| BigCodeBench Complete | 40.5% | — |
| ALE-Bench | — | 599.25 |
| HumanEval+ | 62.8% | — |
| MBPP+ | 55.6% | — |
Agentic & Tool Use MiniMax-M2.7 leads
Llama 3.1-8B: 22.3 (#123), MiniMax-M2.7: 24.9 (#103)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| Terminal-Bench | — | 45.1% |
| Berkeley Function Calling Leaderboard | 25.8% | — |
| BALROG | 15.1% | — |
| ExploitBench | — | 13.3% |
| GBAEval | — | 0% |
Reasoning MiniMax-M2.7 leads
Llama 3.1-8B: 14.5 (#316), MiniMax-M2.7: 19.0 (#248)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| CritPt | 0% | 0.6% |
| LMArena Hard Prompts | 1175 | 1422 |
| Epoch Capabilities Index | 116.48 | 145.84 |
| NYT Connections (extended) | — | 24.7% |
| Chess Puzzles | 0% | — |
| Thematic Generalization | — | 39.3% |
| DTBench | 50.9% | — |
| LMCA | 5.4% | — |
| PIQA | 81.2% | — |
Math MiniMax-M2.7 leads
Llama 3.1-8B: 9.5 (#311), MiniMax-M2.7: 24.9 (#254)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| LMArena Math | 1179 | 1420 |
| OTIS Mock AIME 2024-2025 | 1.7% | — |
| ProofBench | — | 3% |
| Omni-MATH | 13.7% | — |
| MATH Level 5 | 22.9% | — |
| GSM8K | 82.4% | — |
Knowledge MiniMax-M2.7 leads
Llama 3.1-8B: 8.0 (#302), MiniMax-M2.7: 37.1 (#151)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| LMArena Expert | 1144 | 1443 |
| GPQA Diamond | 27% | — |
| MMLU-Pro | 40.6% | — |
| Vectara Hallucination Rate | — | 12.9% |
| GPQA (HELM) | 24.7% | — |
| BoolQ | 82.8% | — |
| MMLU | 56.1% | — |
Multilingual MiniMax-M2.7 leads
Llama 3.1-8B: 34.2 (#244), MiniMax-M2.7: 50.5 (#120)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| LMArena Non-English | 1148 | 1382 |
| LMArena Chinese | 1151 | 1439 |
| LMArena French | 1176 | 1422 |
| LMArena German | 1144 | 1398 |
| LMArena Japanese | 1061 | 1261 |
| LMArena Korean | 1053 | 1313 |
| LMArena Russian | 1158 | 1383 |
| LMArena Spanish | 1169 | 1403 |
Instruction Following MiniMax-M2.7 leads
Llama 3.1-8B: 59.1 (#253), MiniMax-M2.7: 74.2 (#101)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| LMArena Instruction Following | 1159 | 1404 |
| IFEval | 74.3% | — |
Long Context MiniMax-M2.7 leads
Llama 3.1-8B: 35.8 (#233), MiniMax-M2.7: 43.3 (#94)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| LMArena Longer Query | 1182 | 1419 |
Writing & Preference MiniMax-M2.7 leads
Llama 3.1-8B: 29.9 (#285), MiniMax-M2.7: 59.1 (#107)
| Benchmark | Llama 3.1-8B | MiniMax-M2.7 |
|---|---|---|
| LMArena Text | 1187 | 1404 |
| LMArena Creative Writing | 1154 | 1354 |
| LMArena Multi-Turn | 1172 | 1413 |
| EQ-Bench Creative Writing | 713 | — |
| WildBench | 68.7% | — |
Frequently asked questions
Is Llama 3.1-8B better than MiniMax-M2.7?
MiniMax-M2.7 is the stronger model overall, scoring 37.4 to 22.8 on the Noometry Index. Llama 3.1-8B costs 9.1× less per token, which makes it the better buy when MiniMax-M2.7's lead doesn't matter for your workload.
Which is cheaper, Llama 3.1-8B or MiniMax-M2.7?
Llama 3.1-8B is cheaper. It lists at $0.05 per million input tokens and $0.08 per million output tokens; MiniMax-M2.7 lists at $0.30 and $1.20.
Is Llama 3.1-8B or MiniMax-M2.7 better for coding?
MiniMax-M2.7 scores higher on coding benchmarks: 41.4 versus 19.9 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2.7 does, with 205K tokens against 128K.
How many benchmarks do Llama 3.1-8B and MiniMax-M2.7 share?
21 benchmarks have published results for both models. Llama 3.1-8B has 43 scored results on Noometry and MiniMax-M2.7 has 30.