Model comparison
gpt-oss-120b vs Inkling
Inkling is the stronger model overall, scoring 44.1 to 36.3 on the Noometry Index. gpt-oss-120b costs 37× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Last verified . 30 shared benchmarks.
Summary
- They share 30 benchmarks with published results for both. gpt-oss-120b scores higher in 1 category and Inkling in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where gpt-oss-120b leads 52.5 to 31.3.
- The biggest single-benchmark swing is APEX-Agents: 4.4% for gpt-oss-120b and 33.8% for Inkling.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- gpt-oss-120b accepts more context: 131K tokens versus 66K.
Side by side
| gpt-oss-120b | Inkling | |
|---|---|---|
| Provider | OpenAI | Thinking Machines Lab |
| Noometry Index | 36.3 | 44.1 |
| Released | 2025-08-05 | 2026-07-15 |
| Weights | Open | Open |
| Context window | 131K | 66K |
| Max output | 41K | 66K |
| Input $ / M tokens | $0.037 | $1.87 |
| Output $ / M tokens | $0.17 | $4.68 |
| Results tracked | 48 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling leads
gpt-oss-120b: 33.5 (#256), Inkling: 34.5 (#234)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| SciCode | 36% | 47% |
| WeirdML | 48.2% | 32.3% |
| LMArena Coding | 1380 | 1464 |
| ALE-Bench | 575.62 | 946 |
| FrontierCode | — | 14% |
| SWE-bench Verified (bash only) | 26% | — |
| Aider Polyglot | 41.8% | — |
| LMArena WebDev | — | 1413 |
| FrontierSWE | — | 4.1% |
| AlgoTune | 1.41 | — |
Agentic & Tool Use Inkling leads
gpt-oss-120b: 12.2 (#153), Inkling: 29.6 (#85)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| APEX-Agents | 4.4% | 33.8% |
| Terminal-Bench | 18.7% | — |
| τ²-bench Banking | — | 25% |
| METR Time Horizons | 56.6% | — |
| Vending-Bench 2 | -21.53 | — |
Reasoning Inkling leads
gpt-oss-120b: 20.0 (#245), Inkling: 40.4 (#56)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| SimpleBench | 22.1% | 50% |
| CritPt | 1.1% | 5.4% |
| Chess Puzzles | 20% | 21% |
| LMArena Hard Prompts | 1364 | 1451 |
| DTBench | 76.3% | 87.5% |
| LMCA | 22.1% | 37.6% |
| Epoch Capabilities Index | 139.93 | 148.54 |
| ARC-AGI-2 | — | 36.5% |
| Kagi LLM Benchmark | 58.6% | — |
| ARC-AGI-1 | — | 79.5% |
| Mystery Game Puzzles | 2% | — |
| Surface Evolver Bench | 25% | — |
Math gpt-oss-120b leads
gpt-oss-120b: 52.5 (#50), Inkling: 31.3 (#225)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 88.9% |
| LMArena Math | 1389 | 1479 |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| FrontierMath Tier 4 | — | 4.9% |
| ProofBench | — | 0% |
| Omni-MATH | 68.8% | — |
Knowledge Inkling leads
gpt-oss-120b: 42.4 (#96), Inkling: 55.1 (#49)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| GPQA Diamond | 75.8% | 88.3% |
| LMArena Expert | 1356 | 1465 |
| SimpleQA Verified | — | 40.3% |
| MMLU-Pro | 79.5% | — |
| Confabulations | 15.7% | — |
| Vectara Hallucination Rate | 14.2% | — |
| GPQA (HELM) | 68.4% | — |
Multilingual Inkling leads
gpt-oss-120b: 48.0 (#147), Inkling: 54.0 (#52)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| LMArena Non-English | 1351 | 1434 |
| LMArena Chinese | 1385 | 1490 |
| LMArena French | 1369 | 1458 |
| LMArena German | 1353 | 1446 |
| LMArena Japanese | 1331 | 1429 |
| LMArena Korean | 1282 | 1404 |
| LMArena Russian | 1343 | 1429 |
| LMArena Spanish | 1389 | 1448 |
Instruction Following Inkling leads
gpt-oss-120b: 69.3 (#173), Inkling: 75.1 (#71)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| LMArena Instruction Following | 1318 | 1426 |
| IFEval | 83.6% | — |
Long Context Inkling leads
gpt-oss-120b: 31.4 (#278), Inkling: 43.8 (#86)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| LMArena Longer Query | 1319 | 1434 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference Inkling leads
gpt-oss-120b: 46.5 (#217), Inkling: 65.2 (#51)
| Benchmark | gpt-oss-120b | Inkling |
|---|---|---|
| LMArena Text | 1365 | 1441 |
| LMArena Creative Writing | 1275 | 1387 |
| EQ-Bench Creative Writing | 961 | 1611 |
| LMArena Multi-Turn | 1340 | 1436 |
| Short-Story Creative Writing | 77.1% | — |
| WildBench | 84.5% | — |
| EQ-Bench 4 | — | 1226 |
Frequently asked questions
Is gpt-oss-120b better than Inkling?
Inkling is the stronger model overall, scoring 44.1 to 36.3 on the Noometry Index. gpt-oss-120b costs 37× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, gpt-oss-120b or Inkling?
gpt-oss-120b is cheaper. It lists at $0.037 per million input tokens and $0.17 per million output tokens; Inkling lists at $1.87 and $4.68.
Is gpt-oss-120b or Inkling better for coding?
Inkling scores higher on coding benchmarks: 34.5 versus 33.5 in the Noometry coding category.
Which has the bigger context window?
gpt-oss-120b does, with 131K tokens against 66K.
How many benchmarks do gpt-oss-120b and Inkling share?
30 benchmarks have published results for both models. gpt-oss-120b has 48 scored results on Noometry and Inkling has 41.