# Inkling-Small vs Qwen3.8 Max

> Qwen3.8 Max is the stronger model overall, scoring 56.8 to 46.5 on the Noometry Index. Inkling-Small costs 4.7× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/inkling-small-vs-qwen3-8-max
- Last updated: 2026-10-10
- Shared benchmarks: 30

## Summary

- They share 30 benchmarks with published results for both. Inkling-Small scores higher in 1 category and Qwen3.8 Max in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 45.1.
- The biggest single-benchmark swing is ProofBench: 6% for Inkling-Small and 58% for Qwen3.8 Max.
- Inkling-Small is cheaper at $0.45 / $1.20 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- Qwen3.8 Max accepts more context: 1M tokens versus 524K.
- Inkling-Small has downloadable open weights; the other is API-only.

## Snapshot

| | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| Provider | Thinking Machines Lab | Alibaba (Qwen) |
| Noometry Index | 46.5 | 56.8 |
| Rank | 63 | 22 |
| Context | 524K | 1M |
| Input $/M | $0.45 | $2 |
| Output $/M | $1.20 | $6 |
| Weights | Open | Proprietary |

## Coding

- Inkling-Small: 43.6 (#85)
- Qwen3.8 Max: 53.5 (#29)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| LMArena WebDev | 1409 | 1674 |
| SciCode | 48.7% | 53.2% |
| LMArena Coding | 1451 | 1502 |
| DeepSWE | — | 57.5% |
| FrontierSWE | — | 17.8% |

## Agentic & Tool Use

- Inkling-Small: —
- Qwen3.8 Max: 45.4 (#14)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | — | 63.3% |
| τ²-bench Banking | — | 55.1% |
| GDP.pdf | — | 23.2% |

## Reasoning

- Inkling-Small: 38.6 (#63)
- Qwen3.8 Max: 54.4 (#26)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| CritPt | 8.3% | 20% |
| Chess Puzzles | 18% | 40% |
| LMArena Hard Prompts | 1423 | 1496 |
| Mystery Game Puzzles | 6% | 38% |
| Epoch Capabilities Index | 150.15 | 156.41 |
| ARC-AGI-2 | 40.1% | — |
| NYT Connections (extended) | — | 88.3% |
| ARC-AGI-1 | 84% | — |
| DTBench | — | 92% |
| LMCA | — | 46.2% |

## Math

- Inkling-Small: 45.1 (#77)
- Qwen3.8 Max: 73.2 (#20)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 46.3% | 74.7% |
| FrontierMath Tier 4 | 17.1% | 46.3% |
| OTIS Mock AIME 2024-2025 | 90% | 100% |
| ProofBench | 6% | 58% |
| LMArena Math | 1459 | 1499 |

## Knowledge

- Inkling-Small: 48.2 (#77)
- Qwen3.8 Max: 61.7 (#27)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 88.5% | 92.7% |
| SimpleQA Verified | 19.1% | 47.3% |
| LMArena Expert | 1442 | 1507 |

## Multimodal

- Inkling-Small: 39.1 (#62)
- Qwen3.8 Max: 37.2 (#75)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | 1235 | 1314 |
| Furniture Assembly | — | 20% |

## Multilingual

- Inkling-Small: 51.7 (#104)
- Qwen3.8 Max: 56.7 (#18)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1402 | 1472 |
| LMArena Chinese | 1465 | 1538 |
| LMArena French | 1436 | 1503 |
| LMArena German | 1405 | 1483 |
| LMArena Japanese | 1405 | 1467 |
| LMArena Korean | 1363 | 1461 |
| LMArena Russian | 1391 | 1481 |
| LMArena Spanish | 1428 | 1492 |

## Instruction Following

- Inkling-Small: 73.8 (#114)
- Qwen3.8 Max: 77.6 (#17)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1399 | 1479 |

## Long Context

- Inkling-Small: 42.7 (#118)
- Qwen3.8 Max: 45.6 (#31)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1401 | 1489 |

## Writing & Preference

- Inkling-Small: 59.6 (#107)
- Qwen3.8 Max: 67.1 (#30)

| Benchmark | Inkling-Small | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1414 | 1483 |
| LMArena Creative Writing | 1331 | 1479 |
| LMArena Multi-Turn | 1418 | 1489 |
| EQ-Bench Creative Writing | 1491 | — |

## FAQ

### Is Inkling-Small better than Qwen3.8 Max?

Qwen3.8 Max is the stronger model overall, scoring 56.8 to 46.5 on the Noometry Index. Inkling-Small costs 4.7× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.

### Which is cheaper, Inkling-Small or Qwen3.8 Max?

Inkling-Small is cheaper. It lists at $0.45 per million input tokens and $1.20 per million output tokens; Qwen3.8 Max lists at $2 and $6.

### Is Inkling-Small or Qwen3.8 Max better for coding?

Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 43.6 in the Noometry coding category.

### Which has the bigger context window?

Qwen3.8 Max does, with 1M tokens against 524K.

### How many benchmarks do Inkling-Small and Qwen3.8 Max share?

30 benchmarks have published results for both models. Inkling-Small has 33 scored results on Noometry and Qwen3.8 Max has 39.