Model comparison
Inkling-Small vs o4-mini
Inkling-Small is the stronger model overall, scoring 46.5 to 41.6 on the Noometry Index.
Last verified . 29 shared benchmarks.
Summary
- They share 29 benchmarks with published results for both. Inkling-Small scores higher in 6 categories and o4-mini in 3 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling-Small leads 38.6 to 24.6.
- The biggest single-benchmark swing is ARC-AGI-2: 40.1% for Inkling-Small and 6.1% for o4-mini.
- Inkling-Small is cheaper at $0.45 / $1.20 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- Inkling-Small accepts more context: 524K tokens versus 200K.
- Inkling-Small has downloadable open weights; the other is API-only.
Side by side
| Inkling-Small | o4-mini | |
|---|---|---|
| Provider | Thinking Machines Lab | OpenAI |
| Noometry Index | 46.5 | 41.6 |
| Released | 2026-07-15 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 524K | 200K |
| Max output | 1.05M | 100K |
| Input $ / M tokens | $0.45 | $1.10 |
| Output $ / M tokens | $1.20 | $4.40 |
| Results tracked | 33 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling-Small leads
Inkling-Small: 43.6 (#85), o4-mini: 40.9 (#127)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| LMArena Coding | 1451 | 1368 |
| SWE-bench Verified (bash only) | — | 45% |
| Aider Polyglot | — | 72% |
| LMArena WebDev | 1409 | — |
| SciCode | 48.7% | — |
| GSO | — | 3.6% |
| WeirdML | — | 52.6% |
| CadEval | — | 62% |
| ALE-Bench | — | 826.17 |
| AlgoTune | — | 1.72 |
Agentic & Tool Use Not comparable
Inkling-Small: —, o4-mini: 32.6 (#61)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 53.2% |
| GDPval | — | 25.3% |
| METR Time Horizons | — | 63.9% |
Reasoning Inkling-Small leads
Inkling-Small: 38.6 (#63), o4-mini: 24.6 (#162)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| ARC-AGI-2 | 40.1% | 6.1% |
| ARC-AGI-1 | 84% | 58.7% |
| CritPt | 8.3% | 0.6% |
| Chess Puzzles | 18% | 26% |
| LMArena Hard Prompts | 1423 | 1351 |
| Mystery Game Puzzles | 6% | 5% |
| Epoch Capabilities Index | 150.15 | 145.64 |
| SimpleBench | — | 38.7% |
| Kagi LLM Benchmark | — | 67.6% |
| EnigmaEval | — | 9.2% |
| DTBench | — | 77.6% |
| LMCA | — | 26.5% |
| ForecastBench | — | 61.8 |
Math Inkling-Small leads
Inkling-Small: 45.1 (#77), o4-mini: 40.8 (#89)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 46.3% | 36.1% |
| FrontierMath Tier 4 | 17.1% | 4.9% |
| OTIS Mock AIME 2024-2025 | 90% | 81.7% |
| LMArena Math | 1459 | 1389 |
| ProofBench | 6% | — |
| Omni-MATH | — | 72% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge Inkling-Small leads
Inkling-Small: 48.2 (#77), o4-mini: 43.6 (#91)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| GPQA Diamond | 88.5% | 79.6% |
| SimpleQA Verified | 19.1% | 19.6% |
| LMArena Expert | 1442 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| Vectara Hallucination Rate | — | 18.6% |
| GPQA (HELM) | — | 73.5% |
Multimodal o4-mini leads
Inkling-Small: 39.1 (#62), o4-mini: 40.2 (#49)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| LMArena Vision | 1235 | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |
Multilingual Inkling-Small leads
Inkling-Small: 51.7 (#104), o4-mini: 47.0 (#154)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| LMArena Non-English | 1402 | 1337 |
| LMArena Chinese | 1465 | 1354 |
| LMArena French | 1436 | 1364 |
| LMArena German | 1405 | 1336 |
| LMArena Japanese | 1405 | 1308 |
| LMArena Korean | 1363 | 1312 |
| LMArena Russian | 1391 | 1334 |
| LMArena Spanish | 1428 | 1347 |
Instruction Following o4-mini leads
Inkling-Small: 73.8 (#114), o4-mini: 75.2 (#68)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1399 | 1321 |
| IFEval | — | 92.8% |
Long Context o4-mini leads
Inkling-Small: 42.7 (#118), o4-mini: 45.5 (#33)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| LMArena Longer Query | 1401 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference Inkling-Small leads
Inkling-Small: 59.6 (#107), o4-mini: 54.0 (#152)
| Benchmark | Inkling-Small | o4-mini |
|---|---|---|
| LMArena Text | 1414 | 1353 |
| LMArena Creative Writing | 1331 | 1294 |
| LMArena Multi-Turn | 1418 | 1350 |
| Short-Story Creative Writing | — | 75% |
| EQ-Bench Creative Writing | 1491 | — |
| WildBench | — | 85.4% |
Frequently asked questions
Is Inkling-Small better than o4-mini?
Inkling-Small is the stronger model overall, scoring 46.5 to 41.6 on the Noometry Index.
Which is cheaper, Inkling-Small or o4-mini?
Inkling-Small is cheaper. It lists at $0.45 per million input tokens and $1.20 per million output tokens; o4-mini lists at $1.10 and $4.40.
Is Inkling-Small or o4-mini better for coding?
Inkling-Small scores higher on coding benchmarks: 43.6 versus 40.9 in the Noometry coding category.
Which has the bigger context window?
Inkling-Small does, with 524K tokens against 200K.
How many benchmarks do Inkling-Small and o4-mini share?
29 benchmarks have published results for both models. Inkling-Small has 33 scored results on Noometry and o4-mini has 60.