Claude Opus 4.8 vs GLM 5.3: Benchmark Comparison
Claude Opus 4.8 has the higher score on 8 of 17 shared benchmarks; GLM 5.3 leads on 9.
The largest observed score gap is 15.66 pts on Terminal-Bench 4.0 , where GLM 5.3 leads.
Reported ±1 standard-error ranges overlap on 6 of 16 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 Opus 4.8 | GLM 5.3 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals RSI Index | 23.98% | 22.91% | 1.07 pts | Uncertainty comparison unavailable |
| Legal Research Bench | 43.75% ±3.45 | 49.04% ±3.48 | 5.29 pts | Reported ±1 SE ranges overlap |
| LegalBench | 83.57% ±0.47 | 84.84% ±0.40 | 1.27 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 30.00% ±3.30 | 39.70% ±3.52 | 9.70 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 | 53.22% ±2.17 | 42.86% ±2.11 | 10.35 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 85.75% ±1.93 | 88.81% ±2.00 | 3.06 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench-Science | 4.29% ±2.44 | 5.71% ±2.79 | 1.43 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 92.42% ±1.98 | 88.13% ±1.85 | 4.29 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 89.58% ±0.30 | 86.77% ±0.34 | 2.81 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 87.82% ±0.95 | 80.53% ±1.06 | 7.28 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 | 59.23% ±4.50 | 47.51% ±4.52 | 11.72 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 88.60% ±1.42 | 95.40% ±0.94 | 6.80 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 23.23% ±1.34 | 38.89% ±1.82 | 15.66 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 82.72% ±3.08 | 78.13% ±3.94 | 4.60 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 68.13% ±1.21 | 68.54% ±1.21 | 0.41 pts | Reported ±1 SE ranges overlap |
Performance by category
| Category | Claude Opus 4.8 average | GLM 5.3 average |
|---|---|---|
| Legal | 63.66% | 66.94% |
| Finance | 52.82% | 56.03% |
| Healthcare | 69.49% | 65.84% |
| Science | 4.29% | 5.71% |
| Academic | 91.00% | 87.45% |
| Coding | 57.10% | 56.99% |
| Social Mobility | 68.13% | 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 Opus 4.8 cost | GLM 5.3 cost | Claude Opus 4.8 latency | GLM 5.3 latency |
|---|---|---|---|---|
| Vals RSI Index | $465.25 | $61.84 | 90h00m | 90h00m |
| Legal Research Bench | $2.82 | $2.24 | 11m54s | 51m50s |
| LegalBench | N/A | N/A | 3.11s | 22.48s |
| Tax Agent Bench | $1.69 | $1.80 | 7m59s | 50m53s |
| TaxEval v2 | N/A | N/A | 70.33s | 98.32s |
| MedCode | N/A | N/A | 105.92s | 2m49s |
| MedScribe | N/A | N/A | 82.36s | 2m02s |
| Terminal-Bench-Science | $23.14 | $15.63 | 1h55m | 2h45m |
| GPQA Diamond | N/A | N/A | 2m14s | 3m31s |
| MMLU Pro | N/A | N/A | 23.44s | 61.78s |
| LiveCodeBench | N/A | N/A | 2m40s | 4m07s |
| ProgramBench | $31.27 | $21.96 | 1h32m | 4h16m |
| SkillsBench | $4.77 | $0.71 | 15m30s | 13m07s |
| SWE-bench | $1.92 | $0.34 | 9m27s | 14m01s |
| Terminal-Bench 4.0 | $17.14 | $9.37 | 1h11m | 1h23m |
| Vibe Code Bench v1.1 | $26.88 | $12.45 | 1h16m | 1h04m |
| Public Benefits Bench v1.1 | $1.89 | $0.95 | 12m48s | 35m51s |
Results available only for Claude Opus 4.8
- MortgageTax
- MMMU Pro
- 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
- ProofBench v1.1
- MysteryMechanism
- Code Migration v1.1
- IOI
- Vibe Code Bench 1-100
- CyberBench v1.1
- SRE Bench