Grok 4.6 vs GLM 5.3: Benchmark Comparison
Grok 4.6 has the higher score on 11 of 22 shared benchmarks; GLM 5.3 leads on 11.
The largest observed score gap is 21.72 pts on Terminal-Bench 4.0 , where GLM 5.3 leads.
Reported ±1 standard-error ranges overlap on 14 of 21 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 4.6 | GLM 5.3 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals RSI Index | 25.46% | 22.91% | 2.55 pts | Uncertainty comparison unavailable |
| Legal Research Bench | 48.08% ±3.47 | 49.04% ±3.48 | 0.96 pts | Reported ±1 SE ranges overlap |
| LegalBench | 86.31% ±0.42 | 84.84% ±0.40 | 1.47 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 33.14% ±3.40 | 39.70% ±3.52 | 6.56 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 71.10% ±0.90 | 72.36% ±0.88 | 1.27 pts | Reported ±1 SE ranges overlap |
| MedCode | 44.71% ±2.26 | 42.86% ±2.11 | 1.85 pts | Reported ±1 SE ranges overlap |
| MedScribe | 86.53% ±1.96 | 88.81% ±2.00 | 2.28 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 51.00% ±5.02 | 49.00% ±5.02 | 2.00 pts | Reported ±1 SE ranges overlap |
| MysteryMechanism | 30.63% ±3.10 | 22.97% ±2.83 | 7.66 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench-Science | 11.43% ±3.83 | 5.71% ±2.79 | 5.71 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 94.70% ±1.13 | 88.13% ±1.85 | 6.56 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 89.40% ±0.30 | 86.77% ±0.34 | 2.63 pts | Reported ±1 SE ranges do not overlap |
| IOI | 47.61% ±2.09 | 68.44% ±7.48 | 20.83 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 88.22% ±0.94 | 80.53% ±1.06 | 7.69 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 1.00% ±0.70 | 1.50% ±0.86 | 0.50 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 55.77% ±4.84 | 47.51% ±4.52 | 8.26 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 95.60% ±0.92 | 95.40% ±0.94 | 0.20 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 17.17% ±1.34 | 38.89% ±1.82 | 21.72 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 14.75% ±3.17 | 19.99% ±3.81 | 5.24 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 76.24% ±3.82 | 78.13% ±3.94 | 1.89 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 67.68% ±5.87 | 72.08% ±5.41 | 4.41 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 66.85% ±1.23 | 68.54% ±1.21 | 1.69 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Grok 4.6 average | GLM 5.3 average |
|---|---|---|
| Legal | 67.19% | 66.94% |
| Finance | 52.12% | 56.03% |
| Healthcare | 65.62% | 65.84% |
| Math | 51.00% | 49.00% |
| Science | 21.03% | 14.34% |
| Academic | 92.05% | 87.45% |
| Coding | 49.55% | 53.80% |
| Cyber | 67.68% | 72.08% |
| Social Mobility | 66.85% | 68.54% |
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 4.6 cost | GLM 5.3 cost | Grok 4.6 latency | GLM 5.3 latency |
|---|---|---|---|---|
| Vals RSI Index | $88.12 | $61.84 | 90h00m | 90h00m |
| Legal Research Bench | $1.53 | $2.24 | 25m04s | 51m50s |
| LegalBench | N/A | N/A | 27.42s | 22.48s |
| Tax Agent Bench | $0.98 | $1.80 | 10m58s | 50m53s |
| TaxEval v2 | N/A | N/A | 47.57s | 98.32s |
| MedCode | N/A | N/A | 2m07s | 2m49s |
| MedScribe | N/A | N/A | 70.94s | 2m02s |
| ProofBench v1.1 | $0.76 | $2.08 | 14m22s | 42m23s |
| MysteryMechanism | $1.31 | $1.53 | 27m55s | 30m42s |
| Terminal-Bench-Science | $4.82 | $15.63 | 1h18m | 2h45m |
| GPQA Diamond | N/A | N/A | 3m29s | 3m31s |
| MMLU Pro | N/A | N/A | 54.98s | 61.78s |
| IOI | $8.18 | $7.67 | 3h57m | 1h46m |
| LiveCodeBench | N/A | N/A | 118.40s | 4m07s |
| ProgramBench | $16.76 | $21.96 | 4h27m | 4h16m |
| SkillsBench | $0.92 | $0.71 | 13m28s | 13m07s |
| SWE-bench | $0.78 | $0.34 | 10m04s | 14m01s |
| Terminal-Bench 4.0 | $5.06 | $9.37 | 29m25s | 1h23m |
| Vibe Code Bench 1-100 | $26.27 | $8.24 | 2h11m | 1h26m |
| Vibe Code Bench v1.1 | $4.88 | $12.45 | 25m30s | 1h04m |
| CyberBench v1.1 | $4.07 | $2.69 | 36m52s | 30m18s |
| Public Benefits Bench v1.1 | $0.92 | $0.95 | 25m53s | 35m51s |
Results available only for Grok 4.6
- Vals Science Index
- MortgageTax
- BioMysteryBench
- SAGE
- Time Horizon Index: KSP
Results available only for GLM 5.3
- Vals Finance Index
- Vals Index
- Vals Legal Index
- HLAB-Vals-1.1
- EMB v1.1
- Finance Agent v2.1
- Code Migration v1.1
- SRE Bench