Model comparison
GPT-6 Astra vs Inkling
GPT-6 Astra is the stronger model overall, scoring 70.8 to 44.1 on the Noometry Index. Inkling costs 7.8× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Last verified . 38 shared benchmarks.
Summary
- They share 38 benchmarks with published results for both. GPT-6 Astra scores higher in 8 categories and Inkling in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Astra leads 93.5 to 31.3.
- The biggest single-benchmark swing is ProofBench: 99% for GPT-6 Astra and 0% for Inkling.
- Inkling is cheaper at $1.87 / $4.68 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
- GPT-6 Astra accepts more context: 1.05M tokens versus 66K.
- Inkling has downloadable open weights; the other is API-only.
Side by side
| GPT-6 Astra | Inkling | |
|---|---|---|
| Provider | OpenAI | Thinking Machines Lab |
| Noometry Index | 70.8 | 44.1 |
| Released | 2026-09-03 | 2026-07-15 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 66K |
| Max output | 128K | 66K |
| Input $ / M tokens | $10 | $1.87 |
| Output $ / M tokens | $50 | $4.68 |
| Results tracked | 56 | 41 |
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), Inkling: 34.5 (#234)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| FrontierCode | 53.3% | 14% |
| LMArena WebDev | 1786 | 1413 |
| FrontierSWE | 65.5% | 4.1% |
| SciCode | 56.5% | 47% |
| WeirdML | 93.6% | 32.3% |
| LMArena Coding | 1487 | 1464 |
| ALE-Bench | 2,951 | 946 |
| DeepSWE | 74.1% | — |
| GSO | 79.4% | — |
| MirrorCode | 46.7% | — |
Agentic & Tool Use GPT-6 Astra leads
GPT-6 Astra: 52.9 (#3), Inkling: 29.6 (#85)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| APEX-Agents | 64.7% | 33.8% |
| Remote Labor Index | 20.8% | — |
| τ²-bench Banking | — | 25% |
| BALROG | 68.3% | — |
| GDP.pdf | 34.2% | — |
| Vending-Bench 2 | 15,515 | — |
Reasoning GPT-6 Astra leads
GPT-6 Astra: 85.1 (#1), Inkling: 40.4 (#56)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| ARC-AGI-2 | 95% | 36.5% |
| ARC-AGI-1 | 98.5% | 79.5% |
| CritPt | 31.7% | 5.4% |
| Chess Puzzles | 72% | 21% |
| LMArena Hard Prompts | 1462 | 1451 |
| DTBench | 97.3% | 87.5% |
| LMCA | 64.4% | 37.6% |
| Epoch Capabilities Index | 166.45 | 148.54 |
| SimpleBench | — | 50% |
| NYT Connections (extended) | 98.1% | — |
| EBR-Bench | 76.2% | — |
| Mystery Game Puzzles | 84% | — |
| Bench to the Future 3 | 0.14 | — |
Math GPT-6 Astra leads
GPT-6 Astra: 93.5 (#2), Inkling: 31.3 (#225)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| FrontierMath (Tiers 1-3) | 93.7% | 33.3% |
| FrontierMath Tier 4 | 97.6% | 4.9% |
| OTIS Mock AIME 2024-2025 | 100% | 88.9% |
| ProofBench | 99% | 0% |
| LMArena Math | 1465 | 1479 |
| FrontierMath Erdős | 2.9% | — |
Knowledge GPT-6 Astra leads
GPT-6 Astra: 75.3 (#1), Inkling: 55.1 (#49)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| GPQA Diamond | 95.8% | 88.3% |
| SimpleQA Verified | 75.6% | 40.3% |
| LMArena Expert | 1483 | 1465 |
| Humanity's Last Exam | 54.8% | — |
| Vectara Hallucination Rate | 8.7% | — |
Multimodal Not comparable
GPT-6 Astra: 55.0 (#3), Inkling: —
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| LMArena Vision | 1281 | — |
| Blueprint-Bench 2 | 49.7% | — |
| Furniture Assembly | 80% | — |
| LMArena Document | 1468 | — |
Multilingual Too close to call
GPT-6 Astra: 53.7 (#61), Inkling: 54.0 (#52)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| LMArena Non-English | 1430 | 1434 |
| LMArena Chinese | 1484 | 1490 |
| LMArena French | 1456 | 1458 |
| LMArena German | 1440 | 1446 |
| LMArena Japanese | 1379 | 1429 |
| LMArena Korean | 1426 | 1404 |
| LMArena Russian | 1436 | 1429 |
| LMArena Spanish | 1407 | 1448 |
Instruction Following GPT-6 Astra leads
GPT-6 Astra: 76.3 (#44), Inkling: 75.1 (#71)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| LMArena Instruction Following | 1450 | 1426 |
Long Context Too close to call
GPT-6 Astra: 44.5 (#62), Inkling: 43.8 (#86)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| LMArena Longer Query | 1456 | 1434 |
Writing & Preference GPT-6 Astra leads
GPT-6 Astra: 75.3 (#7), Inkling: 65.2 (#51)
| Benchmark | GPT-6 Astra | Inkling |
|---|---|---|
| LMArena Text | 1441 | 1441 |
| LMArena Creative Writing | 1418 | 1387 |
| EQ-Bench Creative Writing | 2173 | 1611 |
| LMArena Multi-Turn | 1448 | 1436 |
| EQ-Bench 4 | — | 1226 |
Frequently asked questions
Is GPT-6 Astra better than Inkling?
GPT-6 Astra is the stronger model overall, scoring 70.8 to 44.1 on the Noometry Index. Inkling costs 7.8× 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 Inkling?
Inkling is cheaper. It lists at $1.87 per million input tokens and $4.68 per million output tokens; GPT-6 Astra lists at $10 and $50.
Is GPT-6 Astra or Inkling better for coding?
GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 34.5 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Astra does, with 1.05M tokens against 66K.
How many benchmarks do GPT-6 Astra and Inkling share?
38 benchmarks have published results for both models. GPT-6 Astra has 56 scored results on Noometry and Inkling has 41.