Model comparison
GPT-6 Astra vs Grok 4.6
GPT-6 Astra is the stronger model overall, scoring 70.8 to 56.9 on the Noometry Index. Grok 4.6 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 . 47 shared benchmarks.
Summary
- They share 47 benchmarks with published results for both. GPT-6 Astra scores higher in 10 categories and Grok 4.6 in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Astra leads 93.5 to 67.0.
- The biggest single-benchmark swing is FrontierMath Tier 4: 97.6% for GPT-6 Astra and 31.7% for Grok 4.6.
- Grok 4.6 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 4.6 | |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 70.8 | 56.9 |
| Released | 2026-09-03 | 2026-08-12 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 500K |
| Max output | 128K | 500K |
| Input $ / M tokens | $10 | $2 |
| Output $ / M tokens | $50 | $6 |
| Results tracked | 56 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Astra leads
GPT-6 Astra: 73.7 (#2), Grok 4.6: 58.5 (#16)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| DeepSWE | 74.1% | 67.5% |
| FrontierCode | 53.3% | 48% |
| LMArena WebDev | 1786 | 1617 |
| FrontierSWE | 65.5% | 25.3% |
| SciCode | 56.5% | 56.5% |
| WeirdML | 93.6% | 67.3% |
| LMArena Coding | 1487 | 1465 |
| ALE-Bench | 2,951 | 1,508 |
| CursorBench | — | 41.4% |
| GSO | 79.4% | — |
| MirrorCode | 46.7% | — |
Agentic & Tool Use GPT-6 Astra leads
GPT-6 Astra: 52.9 (#3), Grok 4.6: 39.4 (#27)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| APEX-Agents | 64.7% | 65.3% |
| GDP.pdf | 34.2% | 17.2% |
| Vending-Bench 2 | 15,515 | 9,047 |
| Remote Labor Index | 20.8% | — |
| BALROG | 68.3% | — |
Reasoning GPT-6 Astra leads
GPT-6 Astra: 85.1 (#1), Grok 4.6: 61.4 (#20)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| ARC-AGI-2 | 95% | 67.1% |
| NYT Connections (extended) | 98.1% | 80% |
| ARC-AGI-1 | 98.5% | 87.5% |
| CritPt | 31.7% | 19.7% |
| Chess Puzzles | 72% | 40% |
| EBR-Bench | 76.2% | 30.5% |
| LMArena Hard Prompts | 1462 | 1447 |
| Mystery Game Puzzles | 84% | 34% |
| DTBench | 97.3% | 97.3% |
| LMCA | 64.4% | 48.5% |
| Epoch Capabilities Index | 166.45 | 156.44 |
| SimpleBench | — | 75.9% |
| Bench to the Future 3 | 0.14 | — |
Math GPT-6 Astra leads
GPT-6 Astra: 93.5 (#2), Grok 4.6: 67.0 (#24)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 93.7% | 66% |
| FrontierMath Tier 4 | 97.6% | 31.7% |
| OTIS Mock AIME 2024-2025 | 100% | 99.2% |
| ProofBench | 99% | 51% |
| LMArena Math | 1465 | 1423 |
| FrontierMath Erdős | 2.9% | — |
Knowledge GPT-6 Astra leads
GPT-6 Astra: 75.3 (#1), Grok 4.6: 63.3 (#20)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| GPQA Diamond | 95.8% | 94% |
| SimpleQA Verified | 75.6% | 49.3% |
| LMArena Expert | 1483 | 1467 |
| Humanity's Last Exam | 54.8% | — |
| Vectara Hallucination Rate | 8.7% | — |
Multimodal GPT-6 Astra leads
GPT-6 Astra: 55.0 (#3), Grok 4.6: 43.6 (#23)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| LMArena Vision | 1281 | 1263 |
| Blueprint-Bench 2 | 49.7% | 33.2% |
| Furniture Assembly | 80% | 40% |
| LMArena Document | 1468 | 1452 |
Multilingual Too close to call
GPT-6 Astra: 53.7 (#61), Grok 4.6: 53.0 (#74)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| LMArena Non-English | 1430 | 1420 |
| LMArena Chinese | 1484 | 1480 |
| LMArena French | 1456 | 1461 |
| LMArena German | 1440 | 1431 |
| LMArena Japanese | 1379 | 1376 |
| LMArena Korean | 1426 | 1397 |
| LMArena Russian | 1436 | 1422 |
| LMArena Spanish | 1407 | 1404 |
Instruction Following Too close to call
GPT-6 Astra: 76.3 (#44), Grok 4.6: 75.4 (#63)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| LMArena Instruction Following | 1450 | 1431 |
Long Context Too close to call
GPT-6 Astra: 44.5 (#62), Grok 4.6: 44.5 (#66)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| LMArena Longer Query | 1456 | 1454 |
Writing & Preference GPT-6 Astra leads
GPT-6 Astra: 75.3 (#7), Grok 4.6: 62.3 (#80)
| Benchmark | GPT-6 Astra | Grok 4.6 |
|---|---|---|
| LMArena Text | 1441 | 1428 |
| LMArena Creative Writing | 1418 | 1428 |
| LMArena Multi-Turn | 1448 | 1425 |
| EQ-Bench Creative Writing | 2173 | — |
Frequently asked questions
Is GPT-6 Astra better than Grok 4.6?
GPT-6 Astra is the stronger model overall, scoring 70.8 to 56.9 on the Noometry Index. Grok 4.6 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 4.6?
Grok 4.6 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 4.6 better for coding?
GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 58.5 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 4.6 share?
47 benchmarks have published results for both models. GPT-6 Astra has 56 scored results on Noometry and Grok 4.6 has 49.