Model comparison
Gemini 3.7 Flash vs GPT-5.6 Terra: Benchmark Comparison
Gemini 3.7 Flash has the higher score on 12 of 26 shared benchmarks; GPT-5.6 Terra leads on 14.
The largest observed score gap is 28.33 pts on CyberBench v1.1 , where GPT-5.6 Terra leads.
Reported ±1 standard-error ranges overlap on 11 of 26 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 3.7 Flash | GPT-5.6 Terra | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 59.31% ±1.06 | 59.59% ±1.33 | 0.28 pts | Reported ±1 SE ranges overlap |
| Harvey's Legal Agent Benchmark | 8.75% ±2.15 | 0.83% ±0.83 | 7.92 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 34.62% ±3.31 | 41.35% ±3.42 | 6.73 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 87.26% ±0.42 | 85.11% ±0.45 | 2.15 pts | Reported ±1 SE ranges do not overlap |
| EMB | 71.33% ±2.26 | 66.20% ±3.11 | 5.12 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 59.04% ±0.27 | 54.44% ±2.07 | 4.61 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 66.65% ±0.92 | 67.33% ±0.93 | 0.67 pts | Reported ±1 SE ranges overlap |
| Tax Agent Bench | 57.66% ±3.31 | 65.20% ±3.19 | 7.53 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 74.73% ±0.85 | 76.17% ±0.84 | 1.43 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.39% ±2.12 | 43.41% ±2.17 | 9.98 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 83.94% ±2.00 | 82.87% ±1.95 | 1.07 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 58.00% ±4.96 | 74.00% ±4.41 | 16.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 5.71% ±2.79 | 10.00% ±3.61 | 4.29 pts | Reported ±1 SE ranges overlap |
| GPQA Diamond | 93.94% ±1.49 | 90.91% ±1.91 | 3.03 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 90.12% ±0.30 | 86.66% ±0.33 | 3.46 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 88.96% ±0.75 | 86.47% ±0.82 | 2.49 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 49.23% ±3.38 | 47.00% ±3.40 | 2.23 pts | Reported ±1 SE ranges overlap |
| Code Migration | 34.80% ±4.22 | 47.80% ±4.28 | 13.00 pts | Reported ±1 SE ranges do not overlap |
| IOI | 67.83% ±3.61 | 87.61% ±6.53 | 19.78 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 88.65% ±0.92 | 85.93% ±1.02 | 2.72 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 0.00% ±0.00 | 0.50% ±0.50 | 0.50 pts | Reported ±1 SE ranges overlap |
| SkillsBench | 65.89% ±4.55 | 58.90% ±4.47 | 7.00 pts | Reported ±1 SE ranges overlap |
| SWE-bench | 80.80% ±1.76 | 95.40% ±0.94 | 14.60 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 6.06% ±0.88 | 26.26% ±0.51 | 20.20 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 70.39% ±4.84 | 74.59% ±4.20 | 4.20 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 |
Performance by category
| Category | Gemini 3.7 Flash average | GPT-5.6 Terra average |
|---|---|---|
| Index | 59.31% | 59.59% |
| Legal | 43.54% | 42.43% |
| Finance | 65.88% | 65.87% |
| Healthcare | 68.67% | 63.14% |
| Math | 58.00% | 74.00% |
| Science | 5.71% | 10.00% |
| Academic | 91.01% | 88.01% |
| Education | 49.23% | 47.00% |
| Coding | 51.80% | 59.62% |
| Beta | 43.75% | 72.08% |
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 3.7 Flash cost | GPT-5.6 Terra cost | Gemini 3.7 Flash latency | GPT-5.6 Terra latency |
|---|---|---|---|---|
| Vals Index | $4.17 | $5.18 | 43m13s | 28m50s |
| Harvey's Legal Agent Benchmark | $2.56 | $3.20 | 7m32s | 17m05s |
| Legal Research Bench | $0.76 | $7.91 | 4m49s | 1h10m |
| LegalBench | N/A | N/A | 2.23s | 2.78s |
| EMB | $5.69 | $2.19 | 10m51s | 13m33s |
| Finance Agent (v2) | $1.48 | $3.63 | 3m01s | 24m10s |
| MortgageTax | N/A | N/A | 4.11s | 6.32s |
| Tax Agent Bench | $0.48 | $4.83 | 87.24s | 1h01m |
| TaxEval v2 | N/A | N/A | 7.83s | 37.28s |
| MedCode | N/A | N/A | 9.16s | 18.41s |
| MedScribe | N/A | N/A | 16.82s | 35.53s |
| ProofBench v1.1 | $0.56 | $1.27 | 7m04s | 12m57s |
| Terminal-Bench Science | $13.22 | $10.58 | 1h08m | 2h47m |
| GPQA Diamond | N/A | N/A | 8.77s | 36.76s |
| MMLU Pro | N/A | N/A | 4.16s | 8.17s |
| MMMU Pro | N/A | N/A | 6.28s | 13.03s |
| SAGE | N/A | N/A | 9.79s | 40.06s |
| Code Migration | $21.46 | $8.13 | 2h39m | 48m34s |
| IOI | $3.40 | $8.66 | 21m09s | 1h23m |
| LiveCodeBench | N/A | N/A | 14.53s | 44.33s |
| ProgramBench | $7.13 | $5.38 | 21m30s | 33m36s |
| SkillsBench | $1.80 | $1.75 | 3m06s | 6m55s |
| SWE-bench | $1.44 | $0.40 | 5m50s | 3m00s |
| Terminal-Bench 4.0 | $14.07 | $5.91 | 1h59m | 1h27m |
| Vibe Code Bench v1.1 | $4.83 | $7.89 | 23m43s | 19m12s |
| CyberBench v1.1 | $2.15 | $3.31 | 7m34s | 17m04s |
Results available only for Gemini 3.7 Flash
- Vals RSI Index
- SRE Bench
Results available only for GPT-5.6 Terra
- Vibe Code Bench 1-100
- Public Benefits Bench v1.1