Gemini 3.8 Flash vs GLM 5.3: Benchmark Comparison
Gemini 3.8 Flash has the higher score on 13 of 30 shared benchmarks; GLM 5.3 leads on 17.
The largest observed score gap is 28.33 pts on CyberBench v1.1 , where GLM 5.3 leads.
Reported ±1 standard-error ranges overlap on 10 of 29 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 | Gemini 3.8 Flash | GLM 5.3 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Finance Index | 60.46% ±1.45 | 60.54% ±1.52 | 0.08 pts | Reported ±1 SE ranges overlap |
| Vals Index | 55.95% ±1.14 | 59.73% ±1.11 | 3.78 pts | Reported ±1 SE ranges do not overlap |
| Vals Legal Index | 69.78% ±1.78 | 74.58% ±1.84 | 4.79 pts | Reported ±1 SE ranges do not overlap |
| Vals RSI Index | 19.93% | 22.91% | 2.98 pts | Uncertainty comparison unavailable |
| HLAB-Vals-1.1 | 12.08% ±2.58 | 14.58% ±2.86 | 2.50 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 38.94% ±3.39 | 49.04% ±3.48 | 10.10 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 86.99% ±0.43 | 84.84% ±0.40 | 2.16 pts | Reported ±1 SE ranges do not overlap |
| EMB v1.1 | 72.87% ±2.42 | 69.11% ±2.65 | 3.76 pts | Reported ±1 SE ranges overlap |
| Finance Agent v2.1 | 65.76% ±2.04 | 65.00% ±2.01 | 0.76 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 32.12% ±3.37 | 39.70% ±3.52 | 7.57 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 74.45% ±0.85 | 72.36% ±0.88 | 2.08 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 48.13% ±2.18 | 42.86% ±2.11 | 5.27 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 84.50% ±1.94 | 88.81% ±2.00 | 4.31 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 48.00% ±5.02 | 49.00% ±5.02 | 1.00 pts | Reported ±1 SE ranges overlap |
| MysteryMechanism | 36.49% ±3.24 | 22.97% ±2.83 | 13.51 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench-Science | 8.57% ±3.37 | 5.71% ±2.79 | 2.86 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 94.44% ±1.48 | 88.13% ±1.85 | 6.31 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 90.22% ±0.29 | 86.77% ±0.34 | 3.45 pts | Reported ±1 SE ranges do not overlap |
| Code Migration v1.1 | 38.65% ±3.65 | 55.07% ±2.12 | 16.43 pts | Reported ±1 SE ranges do not overlap |
| IOI | 56.94% ±2.26 | 68.44% ±7.48 | 11.50 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 89.48% ±0.90 | 80.53% ±1.06 | 8.95 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 | 57.98% ±4.42 | 47.51% ±4.52 | 10.48 pts | Reported ±1 SE ranges do not overlap |
| SWE-bench | 80.00% ±1.79 | 95.40% ±0.94 | 15.40 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 19.70% ±2.31 | 38.89% ±1.82 | 19.19 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 18.77% ±3.31 | 19.99% ±3.81 | 1.22 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 78.65% ±3.88 | 78.13% ±3.94 | 0.53 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 43.75% ±2.21 | 72.08% ±5.41 | 28.33 pts | Reported ±1 SE ranges do not overlap |
| SRE Bench | 11.07% ±1.94 | 9.16% ±1.79 | 1.91 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 65.29% ±1.24 | 68.54% ±1.21 | 3.25 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Gemini 3.8 Flash average | GLM 5.3 average |
|---|---|---|
| Legal | 46.01% | 49.49% |
| Finance | 61.30% | 61.54% |
| Healthcare | 66.32% | 65.84% |
| Math | 48.00% | 49.00% |
| Science | 22.53% | 14.34% |
| Academic | 92.33% | 87.45% |
| Coding | 49.02% | 53.94% |
| Cyber | 27.41% | 40.62% |
| Social Mobility | 65.29% | 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 | Gemini 3.8 Flash cost | GLM 5.3 cost | Gemini 3.8 Flash latency | GLM 5.3 latency |
|---|---|---|---|---|
| Vals Finance Index | $3.70 | $2.21 | 7m44s | 35m44s |
| Vals Index | $6.06 | $8.52 | 35m58s | 1h04m |
| Vals Legal Index | $2.05 | $2.68 | 9m15s | 44m51s |
| Vals RSI Index | $332.04 | $61.84 | 90h00m | 90h00m |
| HLAB-Vals-1.1 | $3.66 | $3.99 | 19m22s | 31m51s |
| Legal Research Bench | $1.63 | $2.24 | 5m24s | 51m50s |
| LegalBench | N/A | N/A | 3.32s | 22.48s |
| EMB v1.1 | $8.24 | $3.79 | 12m49s | 36m12s |
| Finance Agent v2.1 | $1.99 | $1.03 | 7m24s | 20m06s |
| Tax Agent Bench | $0.86 | $1.80 | 2m58s | 50m53s |
| TaxEval v2 | N/A | N/A | 8.76s | 98.32s |
| MedCode | N/A | N/A | 42.89s | 2m49s |
| MedScribe | N/A | N/A | 21.14s | 2m02s |
| ProofBench v1.1 | $0.60 | $2.08 | 6m09s | 42m23s |
| MysteryMechanism | $1.73 | $1.53 | 5m22s | 30m42s |
| Terminal-Bench-Science | $5.64 | $15.63 | 55m11s | 2h45m |
| GPQA Diamond | N/A | N/A | 20.18s | 3m31s |
| MMLU Pro | N/A | N/A | 7.82s | 61.78s |
| Code Migration v1.1 | $15.79 | $34.91 | 2h35m | 2h58m |
| IOI | $3.98 | $7.67 | 14m20s | 1h46m |
| LiveCodeBench | N/A | N/A | 24.07s | 4m07s |
| ProgramBench | $10.93 | $21.96 | 46m39s | 4h16m |
| SkillsBench | $2.34 | $0.71 | 3m55s | 13m07s |
| SWE-bench | $2.19 | $0.34 | 11m23s | 14m01s |
| Terminal-Bench 4.0 | $8.65 | $9.37 | 1h47m | 1h23m |
| Vibe Code Bench 1-100 | $18.91 | $8.24 | 1h06m | 1h26m |
| Vibe Code Bench v1.1 | $6.87 | $12.45 | 8m39s | 1h04m |
| CyberBench v1.1 | $1.31 | $2.69 | 9m49s | 30m18s |
| SRE Bench | $27.46 | $27.47 | 2h28m | 2h48m |
| Public Benefits Bench v1.1 | $1.01 | $0.95 | 3m21s | 35m51s |
Results available only for Gemini 3.8 Flash
- Vals Science Index
- MortgageTax
- BioMysteryBench
- MMMU Pro
- SAGE
- SWE-bench Multimodal
- CUA-bench
- Time Horizon Index: KSP
Results available only for GLM 5.3
None.