Gemini 3.8 Flash vs GPT-5.6 Terra: Benchmark Comparison
Gemini 3.8 Flash has the higher score on 13 of 24 shared benchmarks; GPT-5.6 Terra leads on 11.
The largest observed score gap is 30.67 pts on IOI , where GPT-5.6 Terra leads.
Reported ±1 standard-error ranges overlap on 10 of 24 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 | GPT-5.6 Terra | Gap | Reported uncertainty |
|---|---|---|---|---|
| Legal Research Bench | 38.94% ±3.39 | 41.35% ±3.42 | 2.40 pts | Reported ±1 SE ranges overlap |
| LegalBench | 86.99% ±0.43 | 85.11% ±0.45 | 1.89 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 65.34% ±0.94 | 67.33% ±0.93 | 1.99 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 32.12% ±3.37 | 30.04% ±3.31 | 2.09 pts | Reported ±1 SE ranges overlap |
| TaxEval v2 | 74.45% ±0.85 | 76.17% ±0.84 | 1.72 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 48.13% ±2.18 | 43.41% ±2.17 | 4.72 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 84.50% ±1.94 | 82.87% ±1.95 | 1.63 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 48.00% ±5.02 | 74.00% ±4.41 | 26.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench-Science | 8.57% ±3.37 | 10.00% ±3.61 | 1.43 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 94.44% ±1.48 | 90.91% ±1.91 | 3.53 pts | Reported ±1 SE ranges do not overlap |
| MMLU Pro | 90.22% ±0.29 | 86.66% ±0.33 | 3.56 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 89.08% ±0.75 | 86.47% ±0.82 | 2.60 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 35.06% ±3.36 | 47.00% ±3.40 | 11.94 pts | Reported ±1 SE ranges do not overlap |
| IOI | 56.94% ±2.26 | 87.61% ±6.53 | 30.67 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 89.48% ±0.90 | 85.93% ±1.02 | 3.55 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 1.00% ±0.70 | 0.50% ±0.50 | 0.50 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 57.98% ±4.42 | 58.90% ±4.47 | 0.91 pts | Reported ±1 SE ranges 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 | 22.73% ±2.31 | 3.03 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench 1-100 | 18.77% ±3.31 | 14.82% ±3.12 | 3.95 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 78.65% ±3.88 | 74.59% ±4.20 | 4.06 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 | 9.54% ±1.82 | 1.53 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 65.29% ±1.24 | 62.38% ±1.26 | 2.91 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Gemini 3.8 Flash average | GPT-5.6 Terra average |
|---|---|---|
| Legal | 62.97% | 63.23% |
| Finance | 57.30% | 57.84% |
| Healthcare | 66.32% | 63.14% |
| Math | 48.00% | 74.00% |
| Science | 8.57% | 10.00% |
| Academic | 91.25% | 88.01% |
| Education | 35.06% | 47.00% |
| Coding | 50.32% | 55.06% |
| Cyber | 27.41% | 40.81% |
| Social Mobility | 65.29% | 62.38% |
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 | GPT-5.6 Terra cost | Gemini 3.8 Flash latency | GPT-5.6 Terra latency |
|---|---|---|---|---|
| Legal Research Bench | $1.63 | $7.91 | 5m24s | 1h10m |
| LegalBench | N/A | N/A | 3.32s | 2.78s |
| MortgageTax | N/A | N/A | 14.79s | 6.32s |
| Tax Agent Bench | $0.86 | $4.83 | 2m58s | 1h01m |
| TaxEval v2 | N/A | N/A | 8.76s | 37.28s |
| MedCode | N/A | N/A | 42.89s | 18.41s |
| MedScribe | N/A | N/A | 21.14s | 35.53s |
| ProofBench v1.1 | $0.60 | $1.27 | 6m09s | 12m57s |
| Terminal-Bench-Science | $5.64 | $5.19 | 55m11s | 2h04m |
| GPQA Diamond | N/A | N/A | 20.18s | 36.76s |
| MMLU Pro | N/A | N/A | 7.82s | 8.17s |
| MMMU Pro | N/A | N/A | 12.03s | 13.03s |
| SAGE | N/A | N/A | 28.64s | 40.06s |
| IOI | $3.98 | $8.66 | 14m20s | 1h23m |
| LiveCodeBench | N/A | N/A | 24.07s | 44.33s |
| ProgramBench | $10.93 | $5.38 | 46m39s | 33m36s |
| SkillsBench | $2.34 | $1.75 | 3m55s | 6m55s |
| SWE-bench | $2.19 | $0.40 | 11m23s | 3m00s |
| Terminal-Bench 4.0 | $8.65 | $5.60 | 1h47m | 31m29s |
| Vibe Code Bench 1-100 | $18.91 | $33.33 | 1h06m | 1h49m |
| Vibe Code Bench v1.1 | $6.87 | $7.89 | 8m39s | 19m12s |
| CyberBench v1.1 | $1.31 | $3.31 | 9m49s | 17m04s |
| SRE Bench | $27.46 | $30.86 | 2h28m | 3h17m |
| Public Benefits Bench v1.1 | $1.01 | $1.20 | 3m21s | 19m31s |
Results available only for Gemini 3.8 Flash
- Vals Finance Index
- Vals Index
- Vals Legal Index
- Vals RSI Index
- Vals Science Index
- HLAB-Vals-1.1
- EMB v1.1
- Finance Agent v2.1
- BioMysteryBench
- MysteryMechanism
- Code Migration v1.1
- SWE-bench Multimodal
- CUA-bench
- Time Horizon Index: KSP
Results available only for GPT-5.6 Terra
None.