# Inkling vs Llama 3.1-70B

> Inkling is the stronger model overall, scoring 44.1 to 29.4 on the Noometry Index. Llama 3.1-70B costs 6.4× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/inkling-vs-llama-3-1-70b
- Last updated: 2026-10-10
- Shared benchmarks: 24

## Summary

- They share 24 benchmarks with published results for both. Inkling scores higher in 9 categories and Llama 3.1-70B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Inkling leads 55.2 to 24.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.9% for Inkling and 3.6% for Llama 3.1-70B.
- Llama 3.1-70B is cheaper at $0.40 / $0.40 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Llama 3.1-70B accepts more context: 128K tokens versus 66K.

## Snapshot

| | Inkling | Llama 3.1-70B |
|---|---|---|
| Provider | Thinking Machines Lab | Meta |
| Noometry Index | 44.1 | 29.4 |
| Rank | 75 | 299 |
| Context | 66K | 128K |
| Input $/M | $1.87 | $0.40 |
| Output $/M | $4.68 | $0.40 |
| Weights | Open | Open |

## Coding

- Inkling: 34.1 (#229)
- Llama 3.1-70B: 29.8 (#292)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| WeirdML | 32.3% | 9% |
| LMArena Coding | 1463 | 1260 |
| FrontierCode | 14% | — |
| LMArena WebDev | 1412 | — |
| FrontierSWE | 4.1% | — |
| SciCode | 47% | — |
| BigCodeBench Instruct | — | 46.1% |
| BigCodeBench Complete | — | 54.8% |
| ALE-Bench | 946 | — |

## Agentic & Tool Use

- Inkling: 29.5 (#76)
- Llama 3.1-70B: 24.9 (#104)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| APEX-Agents | 33.8% | — |
| TheAgentCompany | — | 6.9% |
| τ²-bench Banking | 25% | — |
| BALROG | — | 27.9% |

## Reasoning

- Inkling: 40.6 (#48)
- Llama 3.1-70B: 20.8 (#215)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| LMArena Hard Prompts | 1452 | 1241 |
| DTBench | 87.5% | 60% |
| LMCA | 37.6% | 14.8% |
| Epoch Capabilities Index | 148.6 | 125.91 |
| ARC-AGI-2 | 36.5% | — |
| SimpleBench | 50% | — |
| ARC-AGI-1 | 79.5% | — |
| CritPt | 5.4% | — |
| Chess Puzzles | 21% | — |

## Math

- Inkling: 31.2 (#198)
- Llama 3.1-70B: 12.7 (#298)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 3.6% |
| LMArena Math | 1482 | 1252 |
| FrontierMath (Tiers 1-3) | 33.3% | — |
| FrontierMath Tier 4 | 4.9% | — |
| ProofBench | 0% | — |
| Omni-MATH | — | 21% |
| MATH Level 5 | — | 36.7% |

## Knowledge

- Inkling: 55.2 (#44)
- Llama 3.1-70B: 24.1 (#264)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| GPQA Diamond | 88.3% | 44.2% |
| LMArena Expert | 1468 | 1209 |
| SimpleQA Verified | 40.3% | — |
| MMLU-Pro | — | 65.3% |
| GPQA (HELM) | — | 42.6% |
| MMLU | — | 80.1% |

## Multilingual

- Inkling: 54.2 (#49)
- Llama 3.1-70B: 38.9 (#220)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| LMArena Non-English | 1434 | 1219 |
| LMArena Chinese | 1495 | 1215 |
| LMArena French | 1453 | 1260 |
| LMArena German | 1456 | 1222 |
| LMArena Japanese | 1428 | 1132 |
| LMArena Korean | 1401 | 1140 |
| LMArena Russian | 1430 | 1234 |
| LMArena Spanish | 1442 | 1253 |

## Instruction Following

- Inkling: 75.3 (#65)
- Llama 3.1-70B: 65.5 (#218)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| LMArena Instruction Following | 1427 | 1231 |
| IFEval | — | 82.1% |

## Long Context

- Inkling: 43.9 (#79)
- Llama 3.1-70B: 37.6 (#208)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| LMArena Longer Query | 1436 | 1241 |

## Writing & Preference

- Inkling: 65.5 (#48)
- Llama 3.1-70B: 35.6 (#262)

| Benchmark | Inkling | Llama 3.1-70B |
|---|---|---|
| LMArena Text | 1443 | 1261 |
| LMArena Creative Writing | 1390 | 1232 |
| EQ-Bench Creative Writing | 1611 | 784 |
| LMArena Multi-Turn | 1436 | 1256 |
| WildBench | — | 75.8% |
| EQ-Bench 4 | 1226 | — |

## FAQ

### Is Inkling better than Llama 3.1-70B?

Inkling is the stronger model overall, scoring 44.1 to 29.4 on the Noometry Index. Llama 3.1-70B costs 6.4× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.

### Which is cheaper, Inkling or Llama 3.1-70B?

Llama 3.1-70B is cheaper. It lists at $0.40 per million input tokens and $0.40 per million output tokens; Inkling lists at $1.87 and $4.68.

### Is Inkling or Llama 3.1-70B better for coding?

Inkling scores higher on coding benchmarks: 34.1 versus 29.8 in the Noometry coding category.

### Which has the bigger context window?

Llama 3.1-70B does, with 128K tokens against 66K.

### How many benchmarks do Inkling and Llama 3.1-70B share?

24 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and Llama 3.1-70B has 35.