MiniMax-M3 vs GLM 5.3 Flash: Benchmark Comparison
MiniMax-M3 has the higher score on 5 of 19 shared benchmarks; GLM 5.3 Flash leads on 14.
The largest observed score gap is 17.00 pts on SWE-bench , where GLM 5.3 Flash leads.
Reported ±1 standard-error ranges overlap on 6 of 19 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 | GLM 5.3 Flash | Gap | Reported uncertainty |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | 4.17% ±1.65 | 6.67% ±1.65 | 2.50 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 29.81% ±3.18 | 45.19% ±3.46 | 15.38 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 85.42% ±0.44 | 83.93% ±0.43 | 1.49 pts | Reported ±1 SE ranges do not overlap |
| EMB | 47.76% ±2.95 | 55.93% ±3.56 | 8.17 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 48.27% ±0.44 | 57.85% ±2.06 | 9.58 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 49.69% ±3.25 | 61.96% ±3.26 | 12.27 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 72.73% ±0.86 | 75.59% ±0.84 | 2.86 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 87.25% ±1.96 | 88.94% ±1.91 | 1.68 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 18.00% ±3.86 | 21.00% ±4.09 | 3.00 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 92.68% ±1.44 | 86.36% ±2.13 | 6.31 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 84.22% ±0.36 | 86.06% ±0.34 | 1.84 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 81.16% ±0.94 | 86.01% ±0.83 | 4.86 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 19.93% ±3.94 | 20.52% ±4.21 | 0.59 pts | Reported ±1 SE ranges overlap |
| LiveCodeBench | 82.15% ±1.05 | 80.51% ±1.05 | 1.64 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 51.50% ±4.50 | 40.16% ±4.43 | 11.34 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 75.00% ±1.94 | 92.00% ±1.21 | 17.00 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 9.22% ±2.86 | 16.03% ±3.28 | 6.81 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 47.57% ±5.44 | 30.76% ±5.28 | 16.81 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 65.00% ±6.10 | 67.86% ±5.57 | 2.86 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | MiniMax-M3 average | GLM 5.3 Flash average |
|---|---|---|
| Legal | 39.80% | 45.26% |
| Finance | 54.61% | 62.83% |
| Healthcare | 87.25% | 88.94% |
| Math | 18.00% | 21.00% |
| Academic | 86.02% | 86.14% |
| Coding | 47.56% | 46.66% |
| Cyber | 65.00% | 67.86% |
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 | GLM 5.3 Flash cost | MiniMax-M3 latency | GLM 5.3 Flash latency |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | $1.46 | $0.57 | 22m31s | 52m17s |
| Legal Research Bench | $0.34 | $0.12 | 13m34s | 51m07s |
| LegalBench | N/A | N/A | 8.21s | 12.52s |
| EMB | $2.09 | $0.71 | 31m30s | 1h15m |
| Finance Agent (v2) | $0.32 | $0.05 | 8m17s | 14m51s |
| Tax Agent Bench | $0.16 | $0.14 | 5m15s | 21m37s |
| TaxEval v2 | N/A | N/A | 94.04s | 53.06s |
| MedScribe | N/A | N/A | 2m04s | 86.98s |
| ProofBench v1.1 | $0.42 | $0.22 | 11m59s | 1h11m |
| GPQA Diamond | N/A | N/A | 4m39s | 2m11s |
| MMLU Pro | N/A | N/A | 41.39s | 44.76s |
| MMMU Pro | N/A | N/A | 68.35s | 28.22s |
| Code Migration | $7.07 | $3.13 | 1h14m | 3h26m |
| LiveCodeBench | N/A | N/A | 6m07s | 5m02s |
| SkillsBench | $0.50 | $0.04 | 15m28s | 12m09s |
| SWE-bench | $0.42 | $0.02 | 12m06s | 1h04m |
| Vibe Code Bench 1-100 | $3.03 | $0.46 | 26m10s | 39m03s |
| Vibe Code Bench v1.1 | $6.45 | $2.35 | 1h25m | 1h24m |
| CyberBench v1.1 | $3.41 | $0.14 | 40m56s | 35m48s |
Results available only for MiniMax-M3
- Vals Index
- MortgageTax
- MedCode
- SAGE
- Terminal-Bench 4.0
- Public Benefits Bench v1.1
Results available only for GLM 5.3 Flash
- Terminal-Bench Science
- IOI
- ProgramBench