Claude Sonnet 5.5 vs GPT-6 Astra: Benchmark Comparison
Claude Sonnet 5.5 has the higher score on 18 of 24 shared benchmarks; GPT-6 Astra leads on 6.
The largest observed score gap is 26.72 pts on SRE Bench , where GPT-6 Astra 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 | Claude Sonnet 5.5 | GPT-6 Astra | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 70.85% ±0.93 | 65.02% ±1.10 | 5.83 pts | Reported ±1 SE ranges do not overlap |
| Vals Finance Index | 64.53% ±1.48 | 55.64% ±1.38 | 8.89 pts | Reported ±1 SE ranges do not overlap |
| Vals Legal Index | 68.02% ±1.81 | 68.39% ±1.74 | 0.36 pts | Reported ±1 SE ranges overlap |
| Vals Science Index | 59.12% ±2.31 | 65.09% ±2.26 | 5.97 pts | Reported ±1 SE ranges do not overlap |
| HLAB-Vals-1.1 | 10.42% ±2.57 | 7.08% ±1.90 | 3.33 pts | Reported ±1 SE ranges overlap |
| Immigration | 68.96% ±1.54 | 67.69% ±1.55 | 1.27 pts | Reported ±1 SE ranges overlap |
| Legal Research Bench | 48.08% ±3.47 | 39.42% ±3.40 | 8.65 pts | Reported ±1 SE ranges do not overlap |
| EMB v1.1 | 75.90% ±2.47 | 72.68% ±2.30 | 3.22 pts | Reported ±1 SE ranges overlap |
| Finance Agent v2.1 | 67.28% ±2.02 | 60.43% ±2.12 | 6.85 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 41.96% ±3.56 | 20.72% ±2.92 | 21.24 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 52.92% ±2.12 | 48.49% ±2.13 | 4.43 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 91.10% ±1.96 | 87.91% ±1.94 | 3.19 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 100.00% ±0.00 | 99.00% ±1.00 | 1.00 pts | Reported ±1 SE ranges overlap |
| BioMysteryBench | 81.11% ±0.64 | 79.26% ±0.98 | 1.85 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 49.10% ±3.36 | 53.15% ±3.36 | 4.05 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench-Science | 47.14% ±6.01 | 62.86% ±5.82 | 15.71 pts | Reported ±1 SE ranges do not overlap |
| Code Migration v1.1 | 86.88% ±1.52 | 80.58% ±2.34 | 6.30 pts | Reported ±1 SE ranges do not overlap |
| IOI | 83.06% ±3.74 | 100.00% ±0.00 | 16.94 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 6.50% ±1.75 | 5.50% ±1.62 | 1.00 pts | Reported ±1 SE ranges overlap |
| SWE-bench Multimodal | 58.33% ±2.25 | 43.33% ±2.26 | 15.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench 4.0 | 64.14% ±1.01 | 59.60% ±4.40 | 4.55 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 92.39% ±1.26 | 89.59% ±2.17 | 2.80 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 59.58% ±5.21 | 41.07% ±2.56 | 18.51 pts | Reported ±1 SE ranges do not overlap |
| SRE Bench | 30.15% ±2.84 | 56.87% ±3.07 | 26.72 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Claude Sonnet 5.5 average | GPT-6 Astra average |
|---|---|---|
| Legal | 42.49% | 38.07% |
| Finance | 61.71% | 51.28% |
| Healthcare | 72.01% | 68.20% |
| Math | 100.00% | 99.00% |
| Science | 59.12% | 65.09% |
| Coding | 65.22% | 63.10% |
| Cyber | 44.87% | 48.97% |
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 Sonnet 5.5 cost | GPT-6 Astra cost | Claude Sonnet 5.5 latency | GPT-6 Astra latency |
|---|---|---|---|---|
| Vals Index | $16.18 | $21.28 | 1h26m | 33m39s |
| Vals Finance Index | $7.40 | $6.48 | 43m31s | 16m18s |
| Vals Legal Index | $13.37 | $13.94 | 1h15m | 24m08s |
| Vals Science Index | $8.50 | $8.03 | 1h09m | 40m37s |
| HLAB-Vals-1.1 | $17.71 | $25.46 | 1h13m | 32m33s |
| Immigration | $1.16 | $1.37 | 11m03s | 8m21s |
| Legal Research Bench | $13.66 | $10.58 | 1h33m | 24m51s |
| EMB v1.1 | $6.53 | $5.83 | 36m45s | 13m24s |
| Finance Agent v2.1 | $6.95 | $7.82 | 34m26s | 20m29s |
| Tax Agent Bench | $8.73 | $5.80 | 59m21s | 15m01s |
| MedCode | N/A | N/A | 3m59s | 102.86s |
| MedScribe | N/A | N/A | 5m07s | 2m34s |
| ProofBench v1.1 | $0.52 | $1.71 | 5m48s | 5m08s |
| BioMysteryBench | $2.95 | $1.72 | 18m06s | 4m46s |
| MysteryMechanism | $2.98 | $1.56 | 23m30s | 8m09s |
| Terminal-Bench-Science | $19.57 | $20.80 | 2h46m | 1h49m |
| Code Migration v1.1 | $43.88 | $62.07 | 4h07m | 1h27m |
| IOI | $5.81 | $6.50 | 45m51s | 31m40s |
| ProgramBench | $22.62 | $11.57 | 2h07m | 27m02s |
| SWE-bench Multimodal | $3.26 | $1.81 | 24m13s | 7m03s |
| Terminal-Bench 4.0 | $10.75 | $9.58 | 1h22m | 35m42s |
| Vibe Code Bench v1.1 | $19.73 | $38.51 | 1h01m | 39m04s |
| CyberBench v1.1 | $2.91 | $1.84 | 33m53s | 4m10s |
| SRE Bench | $17.27 | $10.24 | 1h55m | 17m51s |
Results available only for Claude Sonnet 5.5
- ClaimBench
- SAGE
- Access to Justice Benchmark
- Public Benefits Bench v1.1
Results available only for GPT-6 Astra
- Vals RSI Index
- Robo-bench
- Vibe Code Bench 1-100
- CUA-bench
- Time Horizon Index: KSP