Claude Fable 5.1 vs GPT-5.6 Luna: Benchmark Comparison
Claude Fable 5.1 has the higher score on 25 of 27 shared benchmarks; GPT-5.6 Luna leads on 2.
The largest observed score gap is 46.47 pts on Terminal-Bench 4.0 , where Claude Fable 5.1 leads.
Reported ±1 standard-error ranges overlap on 6 of 27 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 | GPT-5.6 Luna | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 65.83% ±1.11 | 51.69% ±1.06 | 14.14 pts | Reported ±1 SE ranges do not overlap |
| Harvey's Legal Agent Benchmark | 6.67% ±1.17 | 1.25% ±0.83 | 5.42 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 55.29% ±3.46 | 36.54% ±3.35 | 18.75 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 88.51% ±0.42 | 84.03% ±0.42 | 4.48 pts | Reported ±1 SE ranges do not overlap |
| EMB | 76.67% ±2.08 | 67.12% ±2.94 | 9.55 pts | Reported ±1 SE ranges do not overlap |
| Finance Agent (v2) | 58.88% ±2.06 | 55.04% ±0.31 | 3.83 pts | Reported ±1 SE ranges do not overlap |
| MortgageTax | 70.79% ±0.90 | 67.29% ±0.92 | 3.50 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 77.64% ±2.83 | 60.81% ±3.26 | 16.83 pts | Reported ±1 SE ranges do not overlap |
| TaxEval v2 | 75.96% ±0.83 | 76.17% ±0.84 | 0.20 pts | Reported ±1 SE ranges overlap |
| MedCode | 53.51% ±2.17 | 42.39% ±2.27 | 11.12 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 91.29% ±1.95 | 84.39% ±2.58 | 6.90 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 100.00% ±0.00 | 60.00% ±4.92 | 40.00 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 47.75% ±3.36 | 14.41% ±2.36 | 33.33 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 40.00% ±5.90 | 0.00% ±0.00 | 40.00 pts | Reported ±1 SE ranges do not overlap |
| GPQA Diamond | 93.43% ±1.86 | 91.67% ±1.74 | 1.77 pts | Reported ±1 SE ranges overlap |
| MMLU Pro | 92.38% ±0.27 | 86.04% ±0.35 | 6.34 pts | Reported ±1 SE ranges do not overlap |
| MMMU Pro | 90.64% ±0.70 | 85.03% ±0.86 | 5.61 pts | Reported ±1 SE ranges do not overlap |
| SAGE | 48.53% ±3.34 | 44.22% ±3.33 | 4.31 pts | Reported ±1 SE ranges overlap |
| Code Migration | 54.61% ±4.81 | 44.55% ±4.24 | 10.06 pts | Reported ±1 SE ranges do not overlap |
| IOI | 90.78% ±4.65 | 61.78% ±11.61 | 29.00 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 7.00% ±1.81 | 0.00% ±0.00 | 7.00 pts | Reported ±1 SE ranges do not overlap |
| SkillsBench | 61.55% ±4.76 | 60.45% ±4.75 | 1.10 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 58.08% ±3.31 | 11.62% ±1.01 | 46.47 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench 1-100 | 28.00% ±4.49 | 22.59% ±4.00 | 5.41 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 90.26% ±1.57 | 77.06% ±3.07 | 13.21 pts | Reported ±1 SE ranges do not overlap |
| CyberBench v1.1 | 70.42% ±5.43 | 73.63% ±5.56 | 3.22 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 74.90% ±1.13 | 61.16% ±1.27 | 13.73 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Fable 5.1 average | GPT-5.6 Luna average |
|---|---|---|
| Legal | 50.16% | 40.61% |
| Finance | 71.99% | 65.29% |
| Healthcare | 72.40% | 63.39% |
| Math | 100.00% | 60.00% |
| Science | 43.87% | 7.21% |
| Academic | 92.15% | 87.58% |
| Education | 48.53% | 44.22% |
| Coding | 55.75% | 39.72% |
| Cyber | 70.42% | 73.63% |
| Social Mobility | 74.90% | 61.16% |
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 | GPT-5.6 Luna cost | Claude Fable 5.1 latency | GPT-5.6 Luna latency |
|---|---|---|---|---|
| Vals Index | $28.71 | $0.82 | 1h17m | 29m27s |
| Harvey's Legal Agent Benchmark | $46.21 | $0.38 | 1h52m | 10m56s |
| Legal Research Bench | $23.06 | $0.85 | 57m29s | 39m34s |
| LegalBench | N/A | N/A | 12.40s | 5.89s |
| EMB | $15.93 | $0.37 | 29m32s | 11m00s |
| Finance Agent (v2) | $8.35 | $0.28 | 18m14s | 12m52s |
| MortgageTax | N/A | N/A | 14.98s | 22.84s |
| Tax Agent Bench | $13.18 | $0.47 | 28m34s | 42m46s |
| TaxEval v2 | N/A | N/A | 85.56s | 76.52s |
| MedCode | N/A | N/A | 3m35s | 81.28s |
| MedScribe | N/A | N/A | 3m08s | 116.89s |
| ProofBench v1.1 | $2.70 | $0.12 | 7m58s | 8m34s |
| MysteryMechanism | $5.63 | $0.11 | 17m41s | 8m13s |
| Terminal-Bench Science | $38.01 | $0.58 | 2h36m | 2h03m |
| GPQA Diamond | N/A | N/A | 45.86s | 53.52s |
| MMLU Pro | N/A | N/A | 20.59s | 16.95s |
| MMMU Pro | N/A | N/A | 37.47s | 35.10s |
| SAGE | N/A | N/A | 105.75s | 51.88s |
| Code Migration | $70.97 | $1.88 | 4h20m | 58m35s |
| IOI | $11.20 | $0.58 | 30m06s | 1h10m |
| ProgramBench | $58.52 | $0.94 | 2h27m | 41m34s |
| SkillsBench | $4.52 | $0.21 | 11m02s | 5m20s |
| Terminal-Bench 4.0 | $17.18 | $0.73 | 1h00m | 35m55s |
| Vibe Code Bench 1-100 | $149.66 | $4.91 | 2h53m | 2h03m |
| Vibe Code Bench v1.1 | $33.37 | $0.73 | 57m40s | 25m52s |
| CyberBench v1.1 | $4.01 | $0.40 | 18m32s | 16m15s |
| Public Benefits Bench v1.1 | $7.48 | $0.27 | 48m02s | 38m40s |
Results available only for Claude Fable 5.1
- Vals RSI Index
- LiveCodeBench
- SRE Bench
- CUA-bench
- Time Horizon Index: KSP
Results available only for GPT-5.6 Luna
- BioMysteryBench
- SWE-bench