# Inkling vs Llama 3-70B

> Inkling is the stronger model overall, scoring 44.1 to 28.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/inkling-vs-llama-3-70b
- Last updated: 2026-10-10
- Shared benchmarks: 21

## Summary

- They share 21 benchmarks with published results for both. Inkling scores higher in 8 categories and Llama 3-70B in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Inkling leads 55.2 to 20.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.9% for Inkling and 4.3% for Llama 3-70B.

## Snapshot

| | Inkling | Llama 3-70B |
|---|---|---|
| Provider | Thinking Machines Lab | Meta |
| Noometry Index | 44.1 | 28.5 |
| Rank | 75 | 317 |
| Context | 66K | — |
| Input $/M | $1.87 | — |
| Output $/M | $4.68 | — |
| Weights | Open | Open |

## Coding

- Inkling: 34.1 (#229)
- Llama 3-70B: 35.3 (#215)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| LMArena Coding | 1463 | 1206 |
| FrontierCode | 14% | — |
| LMArena WebDev | 1412 | — |
| FrontierSWE | 4.1% | — |
| SciCode | 47% | — |
| WeirdML | 32.3% | — |
| BigCodeBench Instruct | — | 43.6% |
| BigCodeBench Complete | — | 54.5% |
| ALE-Bench | 946 | — |
| HumanEval+ | — | 72% |
| MBPP+ | — | 69% |

## Agentic & Tool Use

- Inkling: 29.5 (#76)
- Llama 3-70B: 20.9 (#132)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| APEX-Agents | 33.8% | — |
| τ²-bench Banking | 25% | — |
| Cybench | — | 5% |

## Reasoning

- Inkling: 40.6 (#48)
- Llama 3-70B: 17.4 (#284)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| LMArena Hard Prompts | 1452 | 1195 |
| DTBench | 87.5% | 54.2% |
| Epoch Capabilities Index | 148.6 | 122.91 |
| ARC-AGI-2 | 36.5% | — |
| SimpleBench | 50% | — |
| Kagi LLM Benchmark | — | 35.1% |
| ARC-AGI-1 | 79.5% | — |
| CritPt | 5.4% | — |
| Chess Puzzles | 21% | — |
| LMCA | 37.6% | — |
| ForecastBench | — | 57.1 |
| WinoGrande | — | 83.5% |

## Math

- Inkling: 31.2 (#198)
- Llama 3-70B: 12.0 (#299)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 4.3% |
| LMArena Math | 1482 | 1218 |
| FrontierMath (Tiers 1-3) | 33.3% | — |
| FrontierMath Tier 4 | 4.9% | — |
| ProofBench | 0% | — |
| MATH Level 5 | — | 22.6% |

## Knowledge

- Inkling: 55.2 (#44)
- Llama 3-70B: 20.8 (#272)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| GPQA Diamond | 88.3% | 40.6% |
| LMArena Expert | 1468 | 1149 |
| SimpleQA Verified | 40.3% | — |
| MMLU | — | 79.3% |

## Multilingual

- Inkling: 54.2 (#49)
- Llama 3-70B: 33.8 (#246)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| LMArena Non-English | 1434 | 1142 |
| LMArena Chinese | 1495 | 1114 |
| LMArena French | 1453 | 1231 |
| LMArena German | 1456 | 1169 |
| LMArena Japanese | 1428 | 1017 |
| LMArena Korean | 1401 | 1017 |
| LMArena Russian | 1430 | 1159 |
| LMArena Spanish | 1442 | 1241 |

## Instruction Following

- Inkling: 75.3 (#65)
- Llama 3-70B: 62.8 (#233)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| LMArena Instruction Following | 1427 | 1194 |

## Long Context

- Inkling: 43.9 (#79)
- Llama 3-70B: 35.6 (#235)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| LMArena Longer Query | 1436 | 1174 |

## Writing & Preference

- Inkling: 65.5 (#48)
- Llama 3-70B: 43.0 (#225)

| Benchmark | Inkling | Llama 3-70B |
|---|---|---|
| LMArena Text | 1443 | 1221 |
| LMArena Creative Writing | 1390 | 1210 |
| LMArena Multi-Turn | 1436 | 1223 |
| EQ-Bench Creative Writing | 1611 | — |
| EQ-Bench 4 | 1226 | — |

## FAQ

### Is Inkling better than Llama 3-70B?

Inkling is the stronger model overall, scoring 44.1 to 28.5 on the Noometry Index.

### Is Inkling or Llama 3-70B better for coding?

Llama 3-70B scores higher on coding benchmarks: 35.3 versus 34.1 in the Noometry coding category.

### How many benchmarks do Inkling and Llama 3-70B share?

21 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and Llama 3-70B has 31.