Claude Fable 5.1 vs Gemini 3.8 Flash: Benchmark Comparison
Claude Fable 5.1 has the higher score on 28 of 31 shared benchmarks; Gemini 3.8 Flash leads on 3.
The largest observed score gap is 52.00 pts on ProofBench v1.1 , where Claude Fable 5.1 leads.
Reported ±1 standard-error ranges overlap on 6 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 | Claude Fable 5.1 | Gemini 3.8 Flash | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 65.83% ±1.11 | 54.83% ±1.03 | 11.00 pts | Reported ±1 SE ranges do not overlap |
| Vals RSI Index | 36.09% | 19.93% | 16.16 pts | Uncertainty comparison unavailable |
| Harvey's Legal Agent Benchmark | 6.67% ±1.17 | 10.00% ±2.42 | 3.33 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 55.29% ±3.46 | 38.94% ±3.39 | 16.35 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 88.51% ±0.42 | 86.99% ±0.43 | 1.52 pts | Reported ±1 SE ranges do not overlap |
| EMB | 76.67% ±2.08 | 72.20% ±2.42 | 4.47 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 58.88% ±2.06 | 61.44% ±0.13 | 2.56 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 70.79% ±0.90 | 65.34% ±0.94 | 5.44 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 77.64% ±2.83 | 66.77% ±3.15 | 10.87 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 75.96% ±0.83 | 74.45% ±0.85 | 1.51 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.51% ±2.17 | 48.13% ±2.18 | 5.37 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 91.29% ±1.95 | 84.50% ±1.94 | 6.80 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 100.00% ±0.00 | 48.00% ±5.02 | 52.00 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 47.75% ±3.36 | 36.49% ±3.24 | 11.26 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 40.00% ±5.90 | 8.57% ±3.37 | 31.43 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 93.43% ±1.86 | 94.44% ±1.48 | 1.01 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 92.38% ±0.27 | 90.22% ±0.29 | 2.15 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 90.64% ±0.70 | 89.08% ±0.75 | 1.56 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 48.53% ±3.34 | 35.06% ±3.36 | 13.46 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 54.61% ±4.81 | 36.55% ±4.18 | 18.06 pts | Reported ±1 SE ranges do not overlap |
| IOI | 90.78% ±4.65 | 56.94% ±2.26 | 33.83 pts | Reported ±1 SE ranges do not overlap |
| LiveCodeBench | 90.52% ±0.86 | 89.48% ±0.90 | 1.03 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 7.00% ±1.81 | 1.00% ±0.70 | 6.00 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 61.55% ±4.76 | 57.98% ±4.42 | 3.57 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 58.08% ±3.31 | 19.19% ±2.52 | 38.89 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 28.00% ±4.49 | 18.77% ±3.31 | 9.23 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 90.26% ±1.57 | 78.65% ±3.88 | 11.61 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 70.42% ±5.43 | 43.75% ±2.21 | 26.67 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 74.90% ±1.13 | 65.29% ±1.24 | 9.61 pts | Reported ±1 SE ranges do not overlap |
| CUA-bench | 13.17% | 4.17% | 9.00 pts | Uncertainty comparison unavailable |
| Time Horizon Index: KSP | 63.33% ±0.00 | 18.83% ±0.00 | 44.50 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Fable 5.1 average | Gemini 3.8 Flash average |
|---|---|---|
| Legal | 50.16% | 45.31% |
| Finance | 71.99% | 68.04% |
| Healthcare | 72.40% | 66.32% |
| Math | 100.00% | 48.00% |
| Science | 43.87% | 22.53% |
| Academic | 92.15% | 91.25% |
| Education | 48.53% | 35.06% |
| Coding | 60.10% | 44.82% |
| Cyber | 70.42% | 43.75% |
| Social Mobility | 74.90% | 65.29% |
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 Fable 5.1 cost | Gemini 3.8 Flash cost | Claude Fable 5.1 latency | Gemini 3.8 Flash latency |
|---|---|---|---|---|
| Vals Index | $28.71 | $5.73 | 1h17m | 51m57s |
| Vals RSI Index | $298.22 | $332.04 | 90h00m | 90h00m |
| Harvey's Legal Agent Benchmark | $46.21 | $3.66 | 1h52m | 29m21s |
| Legal Research Bench | $23.06 | $1.63 | 57m29s | 5m24s |
| LegalBench | N/A | N/A | 12.40s | 3.32s |
| EMB | $15.93 | $8.24 | 29m32s | 12m49s |
| Finance Agent (v2) | $8.35 | $2.00 | 18m14s | 3m21s |
| MortgageTax | N/A | N/A | 14.98s | 14.79s |
| Tax Agent Bench | $13.18 | $0.86 | 28m34s | 2m58s |
| TaxEval v2 | N/A | N/A | 85.56s | 8.76s |
| MedCode | N/A | N/A | 3m35s | 42.89s |
| MedScribe | N/A | N/A | 3m08s | 21.14s |
| ProofBench v1.1 | $2.70 | $0.60 | 7m58s | 6m09s |
| MysteryMechanism | $5.63 | $1.73 | 17m41s | 5m22s |
| Terminal-Bench Science | $38.01 | $5.64 | 2h36m | 55m11s |
| GPQA Diamond | N/A | N/A | 45.86s | 20.18s |
| MMLU Pro | N/A | N/A | 20.59s | 7.82s |
| MMMU Pro | N/A | N/A | 37.47s | 12.03s |
| SAGE | N/A | N/A | 105.75s | 28.64s |
| Code Migration | $70.97 | $18.49 | 4h20m | 2h37m |
| IOI | $11.20 | $3.98 | 30m06s | 14m20s |
| LiveCodeBench | N/A | N/A | 59.74s | 24.07s |
| ProgramBench | $58.52 | $10.93 | 2h27m | 46m39s |
| SkillsBench | $4.52 | $2.34 | 11m02s | 3m55s |
| Terminal-Bench 4.0 | $17.18 | $8.77 | 1h00m | 1h48m |
| Vibe Code Bench 1-100 | $149.66 | $18.91 | 2h53m | 1h06m |
| Vibe Code Bench v1.1 | $33.37 | $6.87 | 57m40s | 8m39s |
| CyberBench v1.1 | $4.01 | $1.31 | 18m32s | 9m49s |
| Public Benefits Bench v1.1 | $7.48 | $1.01 | 48m02s | 3m21s |
| CUA-bench | $323.78 | $280.03 | 37.44s | 36.55s |
| Time Horizon Index: KSP | $4655.01 | $543.14 | N/A | N/A |
Results available only for Claude Fable 5.1
- SRE Bench
Results available only for Gemini 3.8 Flash
- BioMysteryBench
- SWE-bench