GPT-6 Astra vs Grok 4.5
GPT-6 Astra 
Grok 4.5 Verdict
- GPT-6 Astra (max) leads on quality: 70.6 vs 63.4.
- GPT-6 Astra (max) is stronger in agents & tools, coding, composite, knowledge, maths, reasoning.
- Grok 4.5 is stronger in human preference, long context, multimodal.
- Grok 4.5 is 7.3× cheaper ($3.00 vs $22.00 per 1M blended).
| Metric | GPT-6 Astra (max) | Grok 4.5 |
|---|---|---|
| BenchLeader Index | 70.6 | 63.4 |
| Agents & tools score | 68.2 | 58.6 |
| Coding score | 79.0 | 59.9 |
| Composite score | 73.1 | 66.0 |
| Knowledge score | 79.8 | 76.0 |
| Maths score | 77.0 | – |
| Reasoning score | 70.9 | 68.9 |
| Human preference score | – | 66.9 |
| Long context score | – | 66.2 |
| Multimodal score | – | 64.8 |
| Blended price $/M | $22.00 | $3.00 |
| Output speed | 54 tok/s | 56 tok/s |
| Time to first answer | 328.6 s | 12.7 s |
| Context window | 1.1M | 500k |
| GPQA Diamond | 95.8% | – |
| FrontierMath Tiers 1–3 | 93.7% | – |
| FrontierMath Tier 4 | 97.6% | – |
| OTIS Mock AIME | 100.0% | – |
| SimpleQA Verified | 75.6% | – |
| Terminal-Bench | 58.2% | – |
| SimpleBench | – | 70.0% |
| SciCode | 56.5% | – |
| WeirdML | – | 46.4% |
| APEX-Agents | – | 34.2% |
| FrontierCode | 53.3% | 42.4% |
| Epoch Capabilities Index | – | 153.9 |
| LMArena Text | – | 1471 |
| LMArena Hard Prompts | – | 1495 |
| LMArena Coding | – | 1523 |
| LMArena WebDev | 1796 | 1556 |
| LMArena Vision | – | 1291 |
| LMArena Agent | 12.5 | 3.9 |
| LiveBench | 82.2% | 75.8% |
| LiveBench Reasoning | 92.7% | 87.2% |
| LiveBench Coding | 80.4% | 68.6% |
| LiveBench Agentic Coding | 57.3% | 56.5% |
| LiveBench Mathematics | 96.8% | 90.8% |
| LiveBench Data Analysis | 83.0% | 73.0% |
| LiveBench Language | 89.4% | 82.8% |
| AA Intelligence Index | – | 39.1 |
| AA-LCR | – | 79.3% |
| MMMU-Pro | – | 80.4% |
| AA-Omniscience | – | 25.3 |
| GPQA Diamond (AA) | – | 93.1% |
| Humanity's Last Exam (AA) | – | 42.7% |
| SciCode (AA) | – | 55.0% |
| Terminal-Bench 2.1 (Vals) | 87.3% | – |
| Vals Index | 66.6 | – |
| Kagi LLM Benchmark | – | 83.5% |
| 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.