MiniMax-M3 vs Inkling Small: Benchmark Comparison
MiniMax-M3 has the higher score on 18 of 22 shared benchmarks; Inkling Small leads on 4.
The largest observed score gap is 28.51 pts on Vibe Code Bench v1.1 , where MiniMax-M3 leads.
Reported ±1 standard-error ranges overlap on 5 of 22 shared benchmarks with comparable uncertainty data. This is not a pairwise statistical significance test.
Shared benchmark results
Scores and reported standard errors come from the latest visible Vals benchmark version available for both models.
| Benchmark | MiniMax-M3 | Inkling Small | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Finance Index | 48.37% ±1.41 | 36.27% ±1.56 | 12.09 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 29.81% ±3.18 | 25.48% ±3.03 | 4.33 pts | Reported ±1 SE ranges overlap |
| LegalBench | 85.42% ±0.44 | 82.99% ±0.46 | 2.43 pts | Reported ±1 SE ranges do not overlap |
| EMB v1.1 | 59.75% ±2.54 | 40.82% ±3.21 | 18.93 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent v2.1 | 54.34% ±2.02 | 44.78% ±2.10 | 9.56 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 68.36% ±0.91 | 62.28% ±0.95 | 6.08 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 20.59% ±2.87 | 15.39% ±2.37 | 5.20 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 72.73% ±0.86 | 75.51% ±0.84 | 2.78 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 46.29% ±2.10 | 37.89% ±2.21 | 8.40 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 87.25% ±1.96 | 84.11% ±1.87 | 3.14 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 18.00% ±3.86 | 6.00% ±2.39 | 12.00 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 92.68% ±1.44 | 83.59% ±2.04 | 9.09 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 84.22% ±0.36 | 85.57% ±0.35 | 1.35 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 50.57% ±3.44 | 32.74% ±3.26 | 17.83 pts | Reported ±1 SE ranges do not overlap |
| Code Migration v1.1 | 20.90% ±3.44 | 10.43% ±2.32 | 10.47 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 82.15% ±1.05 | 85.93% ±0.99 | 3.78 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 51.50% ±4.50 | 33.58% ±3.66 | 17.91 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 75.00% ±1.94 | 82.20% ±1.71 | 7.20 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 1.01% ±1.01 | 0.00% ±0.00 | 1.01 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 47.57% ±5.44 | 19.05% ±3.40 | 28.51 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 65.00% ±6.10 | 64.11% ±6.17 | 0.89 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 64.14% ±1.25 | 59.41% ±1.28 | 4.74 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | MiniMax-M3 average | Inkling Small average |
|---|---|---|
| Legal | 57.61% | 54.23% |
| Finance | 55.15% | 47.76% |
| Healthcare | 66.77% | 61.00% |
| Math | 18.00% | 6.00% |
| Academic | 88.45% | 84.58% |
| Education | 50.57% | 32.74% |
| Coding | 46.35% | 38.53% |
| Cyber | 65.00% | 64.11% |
| Social Mobility | 64.14% | 59.41% |
Cost and latency
Cost per test appears only where the benchmark reports it for both models. Latency is the measured completion time for that benchmark.
| Benchmark | MiniMax-M3 cost | Inkling Small cost | MiniMax-M3 latency | Inkling Small latency |
|---|---|---|---|---|
| Vals Finance Index | $1.22 | $0.18 | 10m01s | 10m24s |
| Legal Research Bench | $0.34 | $0.14 | 13m34s | 12m55s |
| LegalBench | N/A | N/A | 8.21s | 17.66s |
| EMB v1.1 | $3.18 | $0.25 | 21m39s | 19m34s |
| Finance Agent v2.1 | $0.30 | $0.21 | 3m09s | 5m14s |
| MortgageTax | N/A | N/A | 26.76s | 112.21s |
| Tax Agent Bench | $0.16 | $0.09 | 5m15s | 6m25s |
| TaxEval v2 | N/A | N/A | 94.04s | 4m02s |
| MedCode | N/A | N/A | 63.12s | 3m11s |
| MedScribe | N/A | N/A | 2m04s | 3m50s |
| ProofBench v1.1 | $0.42 | $0.08 | 11m59s | 8m27s |
| GPQA Diamond | N/A | N/A | 4m39s | 4m20s |
| MMLU Pro | N/A | N/A | 41.39s | 71.69s |
| SAGE | N/A | N/A | 118.54s | 3m26s |
| Code Migration v1.1 | $8.39 | $1.36 | 1h26m | 17m56s |
| LiveCodeBench | N/A | N/A | 6m07s | 5m02s |
| SkillsBench | $0.50 | $0.15 | 15m28s | 11m41s |
| SWE-bench | $0.42 | $0.13 | 12m06s | 4m05s |
| Terminal-Bench 4.0 | $7.23 | $0.32 | 1h16m | 11m45s |
| Vibe Code Bench v1.1 | $6.45 | $0.43 | 1h25m | 49m09s |
| CyberBench v1.1 | $3.41 | $0.25 | 40m56s | 19m35s |
| Public Benefits Bench v1.1 | $0.25 | $0.10 | 11m31s | 8m41s |
Results available only for MiniMax-M3
- Vals Inference Provider Index
- MMMU Pro
- SWE-bench Multimodal
- Vibe Code Bench 1-100
Results available only for Inkling Small
- Vals Index
- Vals Legal Index
- HLAB-Vals-1.1
- IOI
- ProgramBench