GPT-5.4 Pro vs GPT-6 Astra
GPT-5.4 Pro Verdict
- GPT-6 Astra (max) leads on quality: 70.6 vs 64.2.
- GPT-5.4 Pro is stronger in instruction following, multimodal, reasoning.
- GPT-6 Astra (max) is stronger in knowledge, agents & tools, coding, composite, maths.
- GPT-6 Astra (max) is 3.1× cheaper ($22.00 vs $67.50 per 1M blended).
- GPT-6 Astra (max) streams 53.9× faster (54 vs 1 tokens per second).
| Metric | GPT-5.4 Pro | GPT-6 Astra (max) |
|---|---|---|
| BenchLeader Index | 64.2 | 70.6 |
| Instruction following score | 67.3 | – |
| Knowledge score | 73.2 | 79.8 |
| Multimodal score | 69.8 | – |
| Reasoning score | 75.1 | 70.9 |
| Agents & tools score | – | 68.2 |
| Coding score | – | 79.0 |
| Composite score | – | 73.1 |
| Maths score | – | 77.0 |
| Blended price $/M | $67.50 | $22.00 |
| Output speed | 1 tok/s | 54 tok/s |
| Time to first answer | 6.7 s | 328.6 s |
| Context window | 1.1M | 1.1M |
| GPQA Diamond | – | 95.8% |
| FrontierMath Tiers 1–3 | – | 93.7% |
| FrontierMath Tier 4 | – | 97.6% |
| OTIS Mock AIME | – | 100.0% |
| SimpleQA Verified | – | 75.6% |
| Humanity's Last Exam | 44.3% | – |
| Terminal-Bench | – | 58.2% |
| SimpleBench | 74.1% | – |
| SciCode | – | 56.5% |
| FrontierCode | – | 53.3% |
| Epoch Capabilities Index | 158.9 | – |
| LMArena WebDev | – | 1796 |
| LMArena Agent | – | 12.5 |
| LiveBench | – | 82.2% |
| LiveBench Reasoning | – | 92.7% |
| LiveBench Coding | – | 80.4% |
| LiveBench Agentic Coding | – | 57.3% |
| LiveBench Mathematics | – | 96.8% |
| LiveBench Data Analysis | – | 83.0% |
| LiveBench Language | – | 89.4% |
| Terminal-Bench 2.1 (Vals) | – | 87.3% |
| Vals Index | – | 66.6 |
| MultiChallenge | 69.2% | – |
| VISTA | 53.9% | – |
| MultiNRC | 62.3% | – |
| TutorBench | 56.6% | – |
| ARC-AGI-1 | – | 97.5% |
| ARC-AGI-2 | – | 95.0% |
| ARC-AGI-3 | – | 62.7% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.