# Claude Sonnet 4 vs Inkling

> Inkling is the stronger model overall, scoring 44.1 to 40.6 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-sonnet-4-vs-inkling
- Last updated: 2026-10-10
- Shared benchmarks: 30

## Summary

- They share 30 benchmarks with published results for both. Claude Sonnet 4 scores higher in 3 categories and Inkling in 6 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling leads 40.6 to 23.0.
- The biggest single-benchmark swing is ARC-AGI-1: 40% for Claude Sonnet 4 and 79.5% for Inkling.
- Inkling is cheaper at $1.87 / $4.68 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.
- Claude Sonnet 4 accepts more context: 200K tokens versus 66K.
- Inkling has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Sonnet 4 | Inkling |
|---|---|---|
| Provider | Anthropic | Thinking Machines Lab |
| Noometry Index | 40.6 | 44.1 |
| Rank | 139 | 75 |
| Context | 200K | 66K |
| Input $/M | $3 | $1.87 |
| Output $/M | $15 | $4.68 |
| Weights | Proprietary | Open |

## Coding

- Claude Sonnet 4: 44.0 (#70)
- Inkling: 34.1 (#229)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| SciCode | 40% | 47% |
| WeirdML | 46.1% | 32.3% |
| LMArena Coding | 1414 | 1463 |
| ALE-Bench | 655.35 | 946 |
| FrontierCode | — | 14% |
| SWE-bench Verified (bash only) | 64.9% | — |
| Aider Polyglot | 61.3% | — |
| LMArena WebDev | — | 1412 |
| FrontierSWE | — | 4.1% |
| GSO | 4.9% | — |

## Agentic & Tool Use

- Claude Sonnet 4: 38.3 (#29)
- Inkling: 29.5 (#76)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| APEX-Agents | — | 33.8% |
| TheAgentCompany | 33.1% | — |
| τ²-bench Banking | — | 25% |
| Cybench | 35% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
| METR Time Horizons | 62% | — |

## Reasoning

- Claude Sonnet 4: 23.0 (#169)
- Inkling: 40.6 (#48)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| ARC-AGI-2 | 5.9% | 36.5% |
| SimpleBench | 45.5% | 50% |
| ARC-AGI-1 | 40% | 79.5% |
| CritPt | 0.3% | 5.4% |
| LMArena Hard Prompts | 1372 | 1452 |
| DTBench | 77.1% | 87.5% |
| LMCA | 29% | 37.6% |
| Epoch Capabilities Index | 141.69 | 148.6 |
| Kagi LLM Benchmark | 73% | — |
| Chess Puzzles | — | 21% |
| EnigmaEval | 3.1% | — |
| ForecastBench | 60.2 | — |

## Math

- Claude Sonnet 4: 41.5 (#71)
- Inkling: 31.2 (#198)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 88.9% |
| LMArena Math | 1375 | 1482 |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| FrontierMath Tier 4 | — | 4.9% |
| ProofBench | — | 0% |
| Omni-MATH | 60.2% | — |
| MATH Level 5 | 84.4% | — |
| FrontierMath (Feb 2025 set) | 4.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |

## Knowledge

- Claude Sonnet 4: 41.6 (#99)
- Inkling: 55.2 (#44)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| GPQA Diamond | 79.2% | 88.3% |
| LMArena Expert | 1371 | 1468 |
| Humanity's Last Exam | 7.8% | — |
| SimpleQA Verified | — | 40.3% |
| MMLU-Pro | 84.3% | — |
| Confabulations | 13.2% | — |
| Vectara Hallucination Rate | 10.3% | — |
| GPQA (HELM) | 70.6% | — |

## Multimodal

- Claude Sonnet 4: 25.6 (#116)
- Inkling: —

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| LMArena Vision | 1191 | — |
| GeoBench | 37% | — |
| VPCT | 34% | — |
| MindCube | 44.8% | — |

## Multilingual

- Claude Sonnet 4: 46.9 (#151)
- Inkling: 54.2 (#49)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| LMArena Non-English | 1333 | 1434 |
| LMArena Chinese | 1351 | 1495 |
| LMArena French | 1364 | 1453 |
| LMArena German | 1331 | 1456 |
| LMArena Japanese | 1301 | 1428 |
| LMArena Korean | 1293 | 1401 |
| LMArena Russian | 1355 | 1430 |
| LMArena Spanish | 1356 | 1442 |

## Instruction Following

- Claude Sonnet 4: 71.8 (#141)
- Inkling: 75.3 (#65)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| LMArena Instruction Following | 1375 | 1427 |
| IFEval | 84% | — |

## Long Context

- Claude Sonnet 4: 33.6 (#254)
- Inkling: 43.9 (#79)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| LMArena Longer Query | 1398 | 1436 |
| Fiction.LiveBench | 46.9% | — |

## Writing & Preference

- Claude Sonnet 4: 57.3 (#127)
- Inkling: 65.5 (#48)

| Benchmark | Claude Sonnet 4 | Inkling |
|---|---|---|
| LMArena Text | 1351 | 1443 |
| LMArena Creative Writing | 1345 | 1390 |
| EQ-Bench Creative Writing | 1483 | 1611 |
| LMArena Multi-Turn | 1376 | 1436 |
| Short-Story Creative Writing | 81.4% | — |
| WildBench | 83.8% | — |
| EQ-Bench 4 | — | 1226 |

## FAQ

### Is Claude Sonnet 4 better than Inkling?

Inkling is the stronger model overall, scoring 44.1 to 40.6 on the Noometry Index.

### Which is cheaper, Claude Sonnet 4 or Inkling?

Inkling is cheaper. It lists at $1.87 per million input tokens and $4.68 per million output tokens; Claude Sonnet 4 lists at $3 and $15.

### Is Claude Sonnet 4 or Inkling better for coding?

Claude Sonnet 4 scores higher on coding benchmarks: 44.0 versus 34.1 in the Noometry coding category.

### Which has the bigger context window?

Claude Sonnet 4 does, with 200K tokens against 66K.

### How many benchmarks do Claude Sonnet 4 and Inkling share?

30 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and Inkling has 41.