DeepSeek V4.1 Flash vs GLM 5.3: Benchmark Comparison
DeepSeek V4.1 Flash has the higher score on 6 of 21 shared benchmarks; GLM 5.3 leads on 15.
The largest observed score gap is 28.17 pts on IOI , where GLM 5.3 leads.
Reported ±1 standard-error ranges overlap on 13 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 | DeepSeek V4.1 Flash | GLM 5.3 | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 51.32% ±1.13 | 53.51% ±1.30 | 2.19 pts | Reported ±1 SE ranges overlap |
| Harvey's Legal Agent Benchmark | 6.67% ±1.83 | 8.33% ±2.00 | 1.67 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 41.35% ±3.42 | 49.04% ±3.48 | 7.69 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 83.28% ±0.46 | 84.84% ±0.40 | 1.56 pts | Reported ±1 SE ranges do not overlap |
| EMB | 57.21% ±3.23 | 56.34% ±3.32 | 0.86 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 53.48% ±0.39 | 55.84% ±2.07 | 2.36 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 62.46% ±3.15 | 73.09% ±2.97 | 10.63 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 41.17% ±2.04 | 42.86% ±2.11 | 1.69 pts | Reported ±1 SE ranges overlap |
| MedScribe | 85.50% ±1.92 | 88.81% ±2.00 | 3.31 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 54.00% ±5.01 | 49.00% ±5.02 | 5.00 pts | Reported ±1 SE ranges overlap |
| MysteryMechanism | 21.17% ±2.75 | 22.97% ±2.83 | 1.80 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench Science | 0.00% ±0.00 | 5.71% ±2.79 | 5.71 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 45.62% ±4.29 | 44.22% ±4.29 | 1.40 pts | Reported ±1 SE ranges overlap |
| IOI | 40.28% ±2.56 | 68.44% ±7.48 | 28.17 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 0.50% ±0.50 | 1.50% ±0.86 | 1.00 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 69.80% ±3.88 | 47.51% ±4.52 | 22.29 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 19.70% ±1.75 | 38.89% ±1.82 | 19.19 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 16.38% ±3.19 | 19.99% ±3.81 | 3.62 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 84.74% ±2.89 | 78.13% ±3.94 | 6.61 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 73.69% ±5.48 | 72.08% ±5.41 | 1.61 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 64.28% ±1.25 | 68.54% ±1.21 | 4.26 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | DeepSeek V4.1 Flash average | GLM 5.3 average |
|---|---|---|
| Legal | 43.76% | 47.40% |
| Finance | 57.72% | 61.76% |
| Healthcare | 63.34% | 65.84% |
| Math | 54.00% | 49.00% |
| Science | 10.59% | 14.34% |
| Coding | 39.57% | 42.67% |
| Cyber | 73.69% | 72.08% |
| Social Mobility | 64.28% | 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 | DeepSeek V4.1 Flash cost | GLM 5.3 cost | DeepSeek V4.1 Flash latency | GLM 5.3 latency |
|---|---|---|---|---|
| Vals Index | $0.33 | $7.25 | 27m53s | 1h13m |
| Harvey's Legal Agent Benchmark | $0.17 | $4.30 | 9m39s | 42m38s |
| Legal Research Bench | $0.25 | $2.24 | 11m42s | 51m50s |
| LegalBench | N/A | N/A | 6.53s | 22.48s |
| EMB | $0.23 | $3.79 | 13m00s | 36m12s |
| Finance Agent (v2) | $0.21 | $1.07 | 6m04s | 15m50s |
| Tax Agent Bench | $0.13 | $1.80 | 5m49s | 50m53s |
| MedCode | N/A | N/A | 35.13s | 2m49s |
| MedScribe | N/A | N/A | 53.83s | 2m02s |
| ProofBench v1.1 | $0.13 | $2.08 | 9m07s | 42m23s |
| MysteryMechanism | $0.16 | $1.53 | 8m04s | 30m42s |
| Terminal-Bench Science | $0.47 | $15.63 | 2h57m | 2h45m |
| Code Migration | $0.94 | $24.91 | 1h45m | 3h47m |
| IOI | $0.24 | $7.67 | 18m30s | 1h46m |
| ProgramBench | $0.90 | $21.96 | 1h02m | 4h16m |
| SkillsBench | $0.09 | $0.71 | 3m43s | 13m07s |
| Terminal-Bench 4.0 | $0.50 | $9.37 | 48m22s | 1h23m |
| Vibe Code Bench 1-100 | $0.69 | $8.24 | 28m21s | 1h26m |
| Vibe Code Bench v1.1 | $0.41 | $12.45 | 15m28s | 1h04m |
| CyberBench v1.1 | $0.07 | $2.69 | 13m50s | 30m18s |
| Public Benefits Bench v1.1 | $0.07 | $0.95 | 9m59s | 35m51s |
Results available only for DeepSeek V4.1 Flash
- BioMysteryBench
- SAGE
- SRE Bench
Results available only for GLM 5.3
- Vals RSI Index
- TaxEval v2
- GPQA Diamond
- MMLU Pro
- LiveCodeBench
- SWE-bench