Model comparison
GPT-6 Astra vs Grok 4.5
GPT-6 Astra is the stronger model overall, scoring 70.3 to 55.0 on the Noometry Index. Grok 4.5 costs 6.7× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Last verified . 41 shared benchmarks.
Summary
- They share 41 benchmarks with published results for both. GPT-6 Astra scores higher in 8 categories and Grok 4.5 in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Astra leads 93.4 to 60.7.
- The biggest single-benchmark swing is FrontierMath Tier 4: 97.6% for GPT-6 Astra and 24.4% for Grok 4.5.
- Grok 4.5 is cheaper at $2 / $6 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
- GPT-6 Astra accepts more context: 1.05M tokens versus 500K.
Side by side
| GPT-6 Astra | Grok 4.5 | |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 70.3 | 55.0 |
| Released | 2026-09-03 | 2026-07-08 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 500K |
| Max output | 128K | 500K |
| Input $ / M tokens | $10 | $2 |
| Output $ / M tokens | $50 | $6 |
| Results tracked | 51 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Astra leads
GPT-6 Astra: 70.5 (#3), Grok 4.5: 51.7 (#34)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| DeepSWE | 74.1% | 53.8% |
| FrontierCode | 53.3% | 42.4% |
| LMArena WebDev | 1792 | 1552 |
| SciCode | 56.5% | 54.1% |
| WeirdML | 93.6% | 46.4% |
| LMArena Coding | 1488 | 1473 |
| ALE-Bench | 2,951 | 1,309 |
| FrontierSWE | 65.5% | — |
| MirrorCode | 46.7% | — |
Agentic & Tool Use GPT-6 Astra leads
GPT-6 Astra: 52.3 (#3), Grok 4.5: 44.4 (#15)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| APEX-Agents | 64.7% | 56.2% |
| Vending-Bench 2 | 15,515 | 3,887 |
| Remote Labor Index | 20.8% | — |
| τ²-bench Banking | — | 47.9% |
| PostTrainBench | — | 23.4% |
| BALROG | 68.3% | — |
| GBAEval | — | 65.4% |
| GDP.pdf | — | 14% |
| LMArena Search | — | 1213 |
Reasoning GPT-6 Astra leads
GPT-6 Astra: 85.1 (#1), Grok 4.5: 56.0 (#23)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| ARC-AGI-2 | 95% | 52.6% |
| NYT Connections (extended) | 98.1% | 79.9% |
| ARC-AGI-1 | 98.5% | 87.2% |
| CritPt | 31.7% | 15.4% |
| Chess Puzzles | 72% | 36% |
| LMArena Hard Prompts | 1464 | 1462 |
| DTBench | 97.3% | 96.5% |
| LMCA | 64.4% | 45.2% |
| Epoch Capabilities Index | 166.6 | 154.02 |
| SimpleBench | — | 70% |
| Kagi LLM Benchmark | — | 83.5% |
| EBR-Bench | 76.2% | — |
| Mystery Game Puzzles | 84% | — |
| Surface Evolver Bench | — | 74.4% |
| Bench to the Future 3 | 0.14 | — |
Math GPT-6 Astra leads
GPT-6 Astra: 93.4 (#1), Grok 4.5: 60.7 (#29)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 93.7% | 57.2% |
| FrontierMath Tier 4 | 97.6% | 24.4% |
| OTIS Mock AIME 2024-2025 | 100% | 97.8% |
| ProofBench | 99% | 31% |
| LMArena Math | 1463 | 1457 |
| FrontierMath Erdős | 2.9% | — |
Knowledge GPT-6 Astra leads
GPT-6 Astra: 75.3 (#1), Grok 4.5: 62.3 (#21)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| GPQA Diamond | 95.8% | 93.4% |
| SimpleQA Verified | 75.6% | 48.3% |
| LMArena Expert | 1486 | 1466 |
| Humanity's Last Exam | 54.8% | — |
| Vectara Hallucination Rate | 8.7% | — |
Multimodal GPT-6 Astra leads
GPT-6 Astra: 55.1 (#3), Grok 4.5: 37.8 (#63)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| LMArena Vision | 1280 | 1287 |
| Blueprint-Bench 2 | 49.7% | 27.3% |
| Furniture Assembly | 80% | 22.5% |
| LMArena Document | 1468 | 1452 |
Multilingual Too close to call
GPT-6 Astra: 54.0 (#53), Grok 4.5: 54.5 (#43)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| LMArena Non-English | 1431 | 1439 |
| LMArena Chinese | 1495 | 1497 |
| LMArena French | 1467 | 1458 |
| LMArena Russian | 1434 | 1447 |
| LMArena Spanish | 1412 | 1452 |
| LMArena German | — | 1448 |
| LMArena Japanese | — | 1425 |
| LMArena Korean | — | 1402 |
Instruction Following Too close to call
GPT-6 Astra: 76.6 (#37), Grok 4.5: 76.3 (#45)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| LMArena Instruction Following | 1454 | 1446 |
Long Context Too close to call
GPT-6 Astra: 44.7 (#57), Grok 4.5: 44.8 (#54)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| LMArena Longer Query | 1460 | 1463 |
Writing & Preference GPT-6 Astra leads
GPT-6 Astra: 75.7 (#7), Grok 4.5: 65.9 (#40)
| Benchmark | GPT-6 Astra | Grok 4.5 |
|---|---|---|
| LMArena Text | 1443 | 1448 |
| LMArena Creative Writing | 1423 | 1442 |
| EQ-Bench Creative Writing | 2173 | 1579 |
| LMArena Multi-Turn | 1457 | 1457 |
Frequently asked questions
Is GPT-6 Astra better than Grok 4.5?
GPT-6 Astra is the stronger model overall, scoring 70.3 to 55.0 on the Noometry Index. Grok 4.5 costs 6.7× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Which is cheaper, GPT-6 Astra or Grok 4.5?
Grok 4.5 is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; GPT-6 Astra lists at $10 and $50.
Is GPT-6 Astra or Grok 4.5 better for coding?
GPT-6 Astra scores higher on coding benchmarks: 70.5 versus 51.7 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Astra does, with 1.05M tokens against 500K.
How many benchmarks do GPT-6 Astra and Grok 4.5 share?
41 benchmarks have published results for both models. GPT-6 Astra has 51 scored results on Noometry and Grok 4.5 has 52.