Claude Sonnet 5 vs GLM 5.3: Benchmark Comparison
Claude Sonnet 5 has the higher score on 7 of 21 shared benchmarks; GLM 5.3 leads on 13.
The largest observed score gap is 28.28 pts on Terminal-Bench 4.0 , where GLM 5.3 leads.
Reported ±1 standard-error ranges overlap on 7 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 | Claude Sonnet 5 | GLM 5.3 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Legal Research Bench | 41.83% ±3.43 | 49.04% ±3.48 | 7.21 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 83.92% ±0.46 | 84.84% ±0.40 | 0.92 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 29.37% ±3.25 | 39.70% ±3.52 | 10.32 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 75.63% ±0.84 | 72.36% ±0.88 | 3.27 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 47.54% ±2.27 | 42.86% ±2.11 | 4.68 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 76.05% ±3.05 | 88.81% ±2.00 | 12.76 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 77.00% ±4.23 | 49.00% ±5.02 | 28.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench-Science | 5.71% ±2.79 | 5.71% ±2.79 | 0.00 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 88.89% ±2.22 | 88.13% ±1.85 | 0.76 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 87.55% ±0.37 | 86.77% ±0.34 | 0.78 pts | Reported ±1 SE ranges do not overlap |
| IOI | 45.00% ±2.75 | 68.44% ±7.48 | 23.44 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 82.43% ±1.09 | 80.53% ±1.06 | 1.89 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 0.00% ±0.00 | 1.50% ±0.86 | 1.50 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 46.48% ±4.49 | 47.51% ±4.52 | 1.03 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 79.60% ±1.80 | 95.40% ±0.94 | 15.80 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 10.61% ±1.51 | 38.89% ±1.82 | 28.28 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 13.82% ±2.98 | 19.99% ±3.81 | 6.17 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 81.33% ±3.05 | 78.13% ±3.94 | 3.20 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 61.91% ±5.74 | 72.08% ±5.41 | 10.18 pts | Reported ±1 SE ranges overlap |
| SRE Bench | 0.38% ±0.38 | 9.16% ±1.79 | 8.78 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 66.03% ±1.23 | 68.54% ±1.21 | 2.50 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Sonnet 5 average | GLM 5.3 average |
|---|---|---|
| Legal | 62.87% | 66.94% |
| Finance | 52.50% | 56.03% |
| Healthcare | 61.80% | 65.84% |
| Math | 77.00% | 49.00% |
| Science | 5.71% | 5.71% |
| Academic | 88.22% | 87.45% |
| Coding | 44.91% | 53.80% |
| Cyber | 31.14% | 40.62% |
| Social Mobility | 66.03% | 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 | Claude Sonnet 5 cost | GLM 5.3 cost | Claude Sonnet 5 latency | GLM 5.3 latency |
|---|---|---|---|---|
| Legal Research Bench | $2.72 | $2.24 | 25m46s | 51m50s |
| LegalBench | N/A | N/A | 4.90s | 22.48s |
| Tax Agent Bench | $1.79 | $1.80 | 19m19s | 50m53s |
| TaxEval v2 | N/A | N/A | 3m22s | 98.32s |
| MedCode | N/A | N/A | 2m15s | 2m49s |
| MedScribe | N/A | N/A | 4m12s | 2m02s |
| ProofBench v1.1 | $1.37 | $2.08 | 14m58s | 42m23s |
| Terminal-Bench-Science | $30.30 | $15.63 | 3h06m | 2h45m |
| GPQA Diamond | N/A | N/A | 63.23s | 3m31s |
| MMLU Pro | N/A | N/A | 25.19s | 61.78s |
| IOI | $12.89 | $7.67 | 59m44s | 1h46m |
| LiveCodeBench | N/A | N/A | 77.02s | 4m07s |
| ProgramBench | $24.36 | $21.96 | 1h31m | 4h16m |
| SkillsBench | $2.90 | $0.71 | 15m12s | 13m07s |
| SWE-bench | $1.49 | $0.34 | 16m02s | 14m01s |
| Terminal-Bench 4.0 | $26.05 | $9.37 | 1h45m | 1h23m |
| Vibe Code Bench 1-100 | $71.15 | $8.24 | 4h16m | 1h26m |
| Vibe Code Bench v1.1 | $25.39 | $12.45 | 1h07m | 1h04m |
| CyberBench v1.1 | $1.89 | $2.69 | 16m57s | 30m18s |
| SRE Bench | $18.49 | $27.47 | 3h53m | 2h48m |
| Public Benefits Bench v1.1 | $1.29 | $0.95 | 21m09s | 35m51s |
Results available only for Claude Sonnet 5
- MortgageTax
- MMMU Pro
- SAGE
Results available only for GLM 5.3
- Vals Finance Index
- Vals Index
- Vals Legal Index
- Vals RSI Index
- HLAB-Vals-1.1
- EMB v1.1
- Finance Agent v2.1
- MysteryMechanism
- Code Migration v1.1