Model comparison
Inkling vs o4-mini
Inkling is the stronger model overall, scoring 44.1 to 41.6 on the Noometry Index.
Last verified . 32 shared benchmarks.
Summary
- They share 32 benchmarks with published results for both. Inkling scores higher in 4 categories and o4-mini in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling leads 40.4 to 24.6.
- The biggest single-benchmark swing is ARC-AGI-2: 36.5% for Inkling and 6.1% for o4-mini.
- o4-mini is cheaper at $1.10 / $4.40 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- o4-mini accepts more context: 200K tokens versus 66K.
- Inkling has downloadable open weights; the other is API-only.
Side by side
| Inkling | o4-mini | |
|---|---|---|
| Provider | Thinking Machines Lab | OpenAI |
| Noometry Index | 44.1 | 41.6 |
| Released | 2026-07-15 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 66K | 200K |
| Max output | 66K | 100K |
| Input $ / M tokens | $1.87 | $1.10 |
| Output $ / M tokens | $4.68 | $4.40 |
| Results tracked | 41 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o4-mini leads
Inkling: 34.5 (#234), o4-mini: 40.9 (#127)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| WeirdML | 32.3% | 52.6% |
| LMArena Coding | 1464 | 1368 |
| ALE-Bench | 946 | 826.17 |
| FrontierCode | 14% | — |
| SWE-bench Verified (bash only) | — | 45% |
| Aider Polyglot | — | 72% |
| LMArena WebDev | 1413 | — |
| FrontierSWE | 4.1% | — |
| SciCode | 47% | — |
| GSO | — | 3.6% |
| CadEval | — | 62% |
| AlgoTune | — | 1.72 |
Agentic & Tool Use o4-mini leads
Inkling: 29.6 (#85), o4-mini: 32.6 (#61)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| APEX-Agents | 33.8% | — |
| Berkeley Function Calling Leaderboard | — | 53.2% |
| GDPval | — | 25.3% |
| τ²-bench Banking | 25% | — |
| METR Time Horizons | — | 63.9% |
Reasoning Inkling leads
Inkling: 40.4 (#56), o4-mini: 24.6 (#162)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| ARC-AGI-2 | 36.5% | 6.1% |
| SimpleBench | 50% | 38.7% |
| ARC-AGI-1 | 79.5% | 58.7% |
| CritPt | 5.4% | 0.6% |
| Chess Puzzles | 21% | 26% |
| LMArena Hard Prompts | 1451 | 1351 |
| DTBench | 87.5% | 77.6% |
| LMCA | 37.6% | 26.5% |
| Epoch Capabilities Index | 148.54 | 145.64 |
| Kagi LLM Benchmark | — | 67.6% |
| EnigmaEval | — | 9.2% |
| Mystery Game Puzzles | — | 5% |
| ForecastBench | — | 61.8 |
Math o4-mini leads
Inkling: 31.3 (#225), o4-mini: 40.8 (#89)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 33.3% | 36.1% |
| FrontierMath Tier 4 | 4.9% | 4.9% |
| OTIS Mock AIME 2024-2025 | 88.9% | 81.7% |
| LMArena Math | 1479 | 1389 |
| ProofBench | 0% | — |
| Omni-MATH | — | 72% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge Inkling leads
Inkling: 55.1 (#49), o4-mini: 43.6 (#91)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| GPQA Diamond | 88.3% | 79.6% |
| SimpleQA Verified | 40.3% | 19.6% |
| LMArena Expert | 1465 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| Vectara Hallucination Rate | — | 18.6% |
| GPQA (HELM) | — | 73.5% |
Multimodal Not comparable
Inkling: —, o4-mini: 40.2 (#49)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| LMArena Vision | — | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |
Multilingual Inkling leads
Inkling: 54.0 (#52), o4-mini: 47.0 (#154)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| LMArena Non-English | 1434 | 1337 |
| LMArena Chinese | 1490 | 1354 |
| LMArena French | 1458 | 1364 |
| LMArena German | 1446 | 1336 |
| LMArena Japanese | 1429 | 1308 |
| LMArena Korean | 1404 | 1312 |
| LMArena Russian | 1429 | 1334 |
| LMArena Spanish | 1448 | 1347 |
Instruction Following Too close to call
Inkling: 75.1 (#71), o4-mini: 75.2 (#68)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1426 | 1321 |
| IFEval | — | 92.8% |
Long Context o4-mini leads
Inkling: 43.8 (#86), o4-mini: 45.5 (#33)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| LMArena Longer Query | 1434 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference Inkling leads
Inkling: 65.2 (#51), o4-mini: 54.0 (#152)
| Benchmark | Inkling | o4-mini |
|---|---|---|
| LMArena Text | 1441 | 1353 |
| LMArena Creative Writing | 1387 | 1294 |
| LMArena Multi-Turn | 1436 | 1350 |
| Short-Story Creative Writing | — | 75% |
| EQ-Bench Creative Writing | 1611 | — |
| WildBench | — | 85.4% |
| EQ-Bench 4 | 1226 | — |
Frequently asked questions
Is Inkling better than o4-mini?
Inkling is the stronger model overall, scoring 44.1 to 41.6 on the Noometry Index.
Which is cheaper, Inkling or o4-mini?
o4-mini is cheaper. It lists at $1.10 per million input tokens and $4.40 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Inkling or o4-mini better for coding?
o4-mini scores higher on coding benchmarks: 40.9 versus 34.5 in the Noometry coding category.
Which has the bigger context window?
o4-mini does, with 200K tokens against 66K.
How many benchmarks do Inkling and o4-mini share?
32 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and o4-mini has 60.