Model comparison
GPT-6 Luna vs Inkling
GPT-6 Luna is the stronger model overall, scoring 53.3 to 44.1 on the Noometry Index.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. GPT-6 Luna scores higher in 5 categories and Inkling in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Luna leads 76.1 to 31.3.
- The biggest single-benchmark swing is ProofBench: 64% for GPT-6 Luna and 0% for Inkling.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- GPT-6 Luna accepts more context: 1.05M tokens versus 66K.
- Inkling has downloadable open weights; the other is API-only.
Side by side
| GPT-6 Luna | Inkling | |
|---|---|---|
| Provider | OpenAI | Thinking Machines Lab |
| Noometry Index | 53.3 | 44.1 |
| Released | 2026-09-22 | 2026-07-15 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 66K |
| Max output | 128K | 66K |
| Input $ / M tokens | $0.10 | $1.87 |
| Output $ / M tokens | $0.50 | $4.68 |
| Results tracked | 42 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
GPT-6 Luna: 55.5 (#25), Inkling: 34.5 (#234)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| FrontierCode | 42.4% | 14% |
| LMArena WebDev | 1581 | 1413 |
| SciCode | 54.6% | 47% |
| LMArena Coding | 1439 | 1464 |
| ALE-Bench | 1,577 | 946 |
| DeepSWE | 66.6% | — |
| FrontierSWE | — | 4.1% |
| WeirdML | — | 32.3% |
Agentic & Tool Use GPT-6 Luna leads
GPT-6 Luna: 33.3 (#54), Inkling: 29.6 (#85)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| APEX-Agents | 44.3% | 33.8% |
| τ²-bench Banking | — | 25% |
| GDP.pdf | 23% | — |
Reasoning GPT-6 Luna leads
GPT-6 Luna: 48.2 (#41), Inkling: 40.4 (#56)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| ARC-AGI-2 | 59.3% | 36.5% |
| ARC-AGI-1 | 86.7% | 79.5% |
| CritPt | 19.4% | 5.4% |
| Chess Puzzles | 31% | 21% |
| LMArena Hard Prompts | 1411 | 1451 |
| DTBench | 90.1% | 87.5% |
| LMCA | 44.5% | 37.6% |
| Epoch Capabilities Index | 156.28 | 148.54 |
| SimpleBench | — | 50% |
| NYT Connections (extended) | 68.7% | — |
| Mystery Game Puzzles | 7% | — |
Math GPT-6 Luna leads
GPT-6 Luna: 76.1 (#15), Inkling: 31.3 (#225)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| FrontierMath (Tiers 1-3) | 78.9% | 33.3% |
| FrontierMath Tier 4 | 56.1% | 4.9% |
| OTIS Mock AIME 2024-2025 | 98.9% | 88.9% |
| ProofBench | 64% | 0% |
| LMArena Math | 1416 | 1479 |
Knowledge GPT-6 Luna leads
GPT-6 Luna: 57.0 (#41), Inkling: 55.1 (#49)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| GPQA Diamond | 90.5% | 88.3% |
| SimpleQA Verified | 41.4% | 40.3% |
| LMArena Expert | 1444 | 1465 |
Multimodal Not comparable
GPT-6 Luna: 42.4 (#30), Inkling: —
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| LMArena Vision | 1217 | — |
| Blueprint-Bench 2 | 31.2% | — |
| Furniture Assembly | 44.2% | — |
Multilingual Inkling leads
GPT-6 Luna: 50.5 (#117), Inkling: 54.0 (#52)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| LMArena Non-English | 1386 | 1434 |
| LMArena Chinese | 1433 | 1490 |
| LMArena French | 1420 | 1458 |
| LMArena German | 1369 | 1446 |
| LMArena Japanese | 1369 | 1429 |
| LMArena Korean | 1360 | 1404 |
| LMArena Russian | 1394 | 1429 |
| LMArena Spanish | 1393 | 1448 |
Instruction Following Too close to call
GPT-6 Luna: 74.3 (#99), Inkling: 75.1 (#71)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| LMArena Instruction Following | 1409 | 1426 |
Long Context Too close to call
GPT-6 Luna: 43.0 (#111), Inkling: 43.8 (#86)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| LMArena Longer Query | 1409 | 1434 |
Writing & Preference Inkling leads
GPT-6 Luna: 58.3 (#119), Inkling: 65.2 (#51)
| Benchmark | GPT-6 Luna | Inkling |
|---|---|---|
| LMArena Text | 1391 | 1441 |
| LMArena Creative Writing | 1363 | 1387 |
| LMArena Multi-Turn | 1396 | 1436 |
| EQ-Bench Creative Writing | — | 1611 |
| EQ-Bench 4 | — | 1226 |
Frequently asked questions
Is GPT-6 Luna better than Inkling?
GPT-6 Luna is the stronger model overall, scoring 53.3 to 44.1 on the Noometry Index.
Which is cheaper, GPT-6 Luna or Inkling?
GPT-6 Luna is cheaper. It lists at $0.10 per million input tokens and $0.50 per million output tokens; Inkling lists at $1.87 and $4.68.
Is GPT-6 Luna or Inkling better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 34.5 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Luna does, with 1.05M tokens against 66K.
How many benchmarks do GPT-6 Luna and Inkling share?
35 benchmarks have published results for both models. GPT-6 Luna has 42 scored results on Noometry and Inkling has 41.