Grok 4.5 vs GLM 5.2: Benchmark Comparison
Grok 4.5 has the higher score on 12 of 17 shared benchmarks; GLM 5.2 leads on 5.
The largest observed score gap is 20.96 pts on SkillsBench , where Grok 4.5 leads.
Reported ±1 standard-error ranges overlap on 6 of 17 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 | Grok 4.5 | GLM 5.2 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | 12.92% ±2.53 | 7.08% ±2.00 | 5.83 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 37.98% ±3.37 | 31.25% ±3.22 | 6.73 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 85.97% ±0.41 | 84.07% ±0.45 | 1.89 pts | Reported ±1 SE ranges do not overlap |
| EMB | 52.91% ±3.11 | 61.53% ±2.88 | 8.62 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 48.35% ±0.46 | 49.70% ±0.88 | 1.35 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 71.67% ±0.89 | 73.34% ±0.86 | 1.68 pts | Reported ±1 SE ranges overlap |
| MedCode | 43.29% ±2.31 | 40.77% ±2.17 | 2.52 pts | Reported ±1 SE ranges overlap |
| MedScribe | 86.88% ±1.94 | 83.53% ±2.00 | 3.35 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 92.93% ±1.29 | 85.61% ±1.77 | 7.32 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 89.22% ±0.31 | 86.71% ±0.33 | 2.50 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 36.60% ±4.14 | 37.87% ±4.14 | 1.27 pts | Reported ±1 SE ranges overlap |
| LiveCodeBench | 87.35% ±0.96 | 69.50% ±1.18 | 17.85 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 0.00% ±0.00 | 0.50% ±0.50 | 0.50 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 66.03% ±4.60 | 45.08% ±4.39 | 20.96 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 86.60% ±1.52 | 82.80% ±1.69 | 3.80 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 69.00% ±4.48 | 63.96% ±4.79 | 5.04 pts | Reported ±1 SE ranges overlap |
| SRE Bench | 0.76% ±0.54 | 0.00% ±0.00 | 0.76 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Grok 4.5 average | GLM 5.2 average |
|---|---|---|
| Legal | 45.62% | 40.80% |
| Finance | 57.64% | 61.52% |
| Healthcare | 65.09% | 62.15% |
| Academic | 91.07% | 86.16% |
| Coding | 57.60% | 49.95% |
| Cyber | 0.76% | 0.00% |
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 | Grok 4.5 cost | GLM 5.2 cost | Grok 4.5 latency | GLM 5.2 latency |
|---|---|---|---|---|
| Harvey's Legal Agent Benchmark | $2.02 | $2.06 | 10m15s | 20m51s |
| Legal Research Bench | $1.09 | $0.89 | 15m54s | 17m03s |
| LegalBench | N/A | N/A | 67.88s | 5.77s |
| EMB | $1.55 | $4.47 | 18m28s | 58m52s |
| Finance Agent (v2) | $1.16 | $0.71 | 4m49s | 7m53s |
| TaxEval v2 | N/A | N/A | 3m15s | 64.94s |
| MedCode | N/A | N/A | 59.64s | 95.27s |
| MedScribe | N/A | N/A | 11m45s | 2m18s |
| GPQA Diamond | N/A | N/A | 3m38s | 3m45s |
| MMLU Pro | N/A | N/A | 5m58s | 51.64s |
| Code Migration | $3.37 | $12.64 | 23m29s | 1h54m |
| LiveCodeBench | N/A | N/A | 9m49s | 8m52s |
| ProgramBench | $4.12 | $12.58 | 18m59s | 2h06m |
| SkillsBench | $0.58 | $0.81 | 2m16s | 19m52s |
| SWE-bench | $0.54 | $0.71 | 3m20s | 11m10s |
| Vibe Code Bench v1.1 | $4.10 | $8.49 | 13m50s | 1h02m |
| SRE Bench | $13.42 | $26.58 | 1h06m | 2h29m |
Results available only for Grok 4.5
- Vals Index
- MortgageTax
- Tax Agent Bench
- ProofBench v1.1
- MMMU Pro
- SAGE
- IOI
- Terminal-Bench 4.0
- CyberBench v1.1
- Time Horizon Index: KSP
Results available only for GLM 5.2
None.