Model comparison
Inkling-Small vs Trinity Large Thinking
Inkling-Small is the stronger model overall, scoring 46.5 to 38.1 on the Noometry Index. Trinity Large Thinking costs 1.6× less per token, which makes it the better buy when Inkling-Small's lead doesn't matter for your workload.
Last verified . 20 shared benchmarks.
Arcee AISummary
- They share 20 benchmarks with published results for both. Inkling-Small scores higher in 8 categories and Trinity Large Thinking in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling-Small leads 39.0 to 16.3.
- The biggest single-benchmark swing is SciCode: 48.7% for Inkling-Small and 36.1% for Trinity Large Thinking.
- Trinity Large Thinking is cheaper at $0.25 / $0.80 per million input/output tokens, against $0.45 / $1.20 for Inkling-Small.
- Inkling-Small accepts more context: 524K tokens versus 262K.
Side by side
| Inkling-Small | Trinity Large Thinking | |
|---|---|---|
| Provider | Thinking Machines Lab | Arcee AI |
| Noometry Index | 46.5 | 38.1 |
| Released | 2026-07-15 | 2026-04-01 |
| Weights | Open | Open |
| Context window | 524K | 262K |
| Max output | 1.05M | 80K |
| Input $ / M tokens | $0.45 | $0.25 |
| Output $ / M tokens | $1.20 | $0.80 |
| Results tracked | 33 | 24 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling-Small leads
Inkling-Small: 43.3 (#80), Trinity Large Thinking: 33.8 (#238)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena WebDev | 1408 | 1238 |
| SciCode | 48.7% | 36.1% |
| LMArena Coding | 1451 | 1381 |
Reasoning Inkling-Small leads
Inkling-Small: 39.0 (#53), Trinity Large Thinking: 16.3 (#298)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| CritPt | 8.3% | 0.9% |
| LMArena Hard Prompts | 1424 | 1350 |
| ARC-AGI-2 | 40.1% | — |
| NYT Connections (extended) | — | 16.5% |
| ARC-AGI-1 | 84% | — |
| Chess Puzzles | 18% | — |
| Thematic Generalization | — | 41.6% |
| Mystery Game Puzzles | 6% | — |
| Surface Evolver Bench | — | 15.6% |
| Epoch Capabilities Index | 150.13 | — |
Math Inkling-Small leads
Inkling-Small: 44.9 (#65), Trinity Large Thinking: 34.9 (#149)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena Math | 1453 | 1366 |
| FrontierMath (Tiers 1-3) | 46.3% | — |
| FrontierMath Tier 4 | 17.1% | — |
| OTIS Mock AIME 2024-2025 | 90% | — |
| ProofBench | 6% | — |
Knowledge Inkling-Small leads
Inkling-Small: 48.3 (#71), Trinity Large Thinking: 40.3 (#107)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena Expert | 1442 | 1359 |
| GPQA Diamond | 88.5% | — |
| SimpleQA Verified | 19.1% | — |
| Vectara Hallucination Rate | — | 6.9% |
Multimodal Not comparable
Inkling-Small: 38.4 (#59), Trinity Large Thinking: —
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena Vision | 1235 | — |
Multilingual Inkling-Small leads
Inkling-Small: 51.7 (#102), Trinity Large Thinking: 46.4 (#155)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena Non-English | 1400 | 1325 |
| LMArena Chinese | 1458 | 1372 |
| LMArena French | 1438 | 1374 |
| LMArena German | 1412 | 1356 |
| LMArena Japanese | 1399 | 1312 |
| LMArena Korean | 1355 | 1304 |
| LMArena Russian | 1392 | 1338 |
| LMArena Spanish | 1426 | 1356 |
Instruction Following Inkling-Small leads
Inkling-Small: 73.9 (#112), Trinity Large Thinking: 70.7 (#157)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena Instruction Following | 1397 | 1334 |
Long Context Inkling-Small leads
Inkling-Small: 42.7 (#113), Trinity Large Thinking: 41.2 (#139)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena Longer Query | 1401 | 1355 |
Writing & Preference Inkling-Small leads
Inkling-Small: 59.8 (#102), Trinity Large Thinking: 54.0 (#152)
| Benchmark | Inkling-Small | Trinity Large Thinking |
|---|---|---|
| LMArena Text | 1413 | 1340 |
| LMArena Creative Writing | 1330 | 1321 |
| LMArena Multi-Turn | 1420 | 1342 |
| EQ-Bench Creative Writing | 1491 | — |
Frequently asked questions
Is Inkling-Small better than Trinity Large Thinking?
Inkling-Small is the stronger model overall, scoring 46.5 to 38.1 on the Noometry Index. Trinity Large Thinking costs 1.6× less per token, which makes it the better buy when Inkling-Small's lead doesn't matter for your workload.
Which is cheaper, Inkling-Small or Trinity Large Thinking?
Trinity Large Thinking is cheaper. It lists at $0.25 per million input tokens and $0.80 per million output tokens; Inkling-Small lists at $0.45 and $1.20.
Is Inkling-Small or Trinity Large Thinking better for coding?
Inkling-Small scores higher on coding benchmarks: 43.3 versus 33.8 in the Noometry coding category.
Which has the bigger context window?
Inkling-Small does, with 524K tokens against 262K.
How many benchmarks do Inkling-Small and Trinity Large Thinking share?
20 benchmarks have published results for both models. Inkling-Small has 33 scored results on Noometry and Trinity Large Thinking has 24.