# GPT-5.4 mini vs Llama 3-70B

> GPT-5.4 mini is the stronger model overall, scoring 44.8 to 28.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-5-4-mini-vs-llama-3-70b
- Last updated: 2026-10-10
- Shared benchmarks: 23

## Summary

- They share 23 benchmarks with published results for both. GPT-5.4 mini scores higher in 9 categories and Llama 3-70B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 mini leads 45.4 to 12.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.9% for GPT-5.4 mini and 4.3% for Llama 3-70B.
- Llama 3-70B has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 44.8 | 28.5 |
| Rank | 69 | 317 |
| Context | 400K | — |
| Input $/M | $0.75 | — |
| Output $/M | $4.50 | — |
| Weights | Proprietary | Open |

## Coding

- GPT-5.4 mini: 45.0 (#67)
- Llama 3-70B: 35.3 (#215)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| LMArena Coding | 1438 | 1206 |
| FrontierCode | 27% | — |
| LMArena WebDev | 1397 | — |
| SciCode | 49.9% | — |
| WeirdML | 60.3% | — |
| BigCodeBench Instruct | — | 43.6% |
| BigCodeBench Complete | — | 54.5% |
| ALE-Bench | 1,189 | — |
| HumanEval+ | — | 72% |
| MBPP+ | — | 69% |

## Agentic & Tool Use

- GPT-5.4 mini: 29.6 (#74)
- Llama 3-70B: 20.9 (#132)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| Cybench | — | 5% |
| DeepResearch Bench | 36.3% | — |

## Reasoning

- GPT-5.4 mini: 30.3 (#77)
- Llama 3-70B: 17.4 (#284)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| Kagi LLM Benchmark | 37.9% | 35.1% |
| LMArena Hard Prompts | 1424 | 1195 |
| DTBench | 80% | 54.2% |
| Epoch Capabilities Index | 148.83 | 122.91 |
| ForecastBench | 57 | 57.1 |
| ARC-AGI-2 | 18.9% | — |
| NYT Connections (extended) | 61.8% | — |
| ARC-AGI-1 | 63.7% | — |
| CritPt | 10% | — |
| Chess Puzzles | 24% | — |
| Thematic Generalization | 61.7% | — |
| Mystery Game Puzzles | 11% | — |
| LMCA | 40.8% | — |
| WinoGrande | — | 83.5% |

## Math

- GPT-5.4 mini: 45.4 (#63)
- Llama 3-70B: 12.0 (#299)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 4.3% |
| LMArena Math | 1418 | 1218 |
| FrontierMath (Tiers 1-3) | 51.2% | — |
| FrontierMath Tier 4 | 9.8% | — |
| ProofBench | 21% | — |
| MATH Level 5 | — | 22.6% |
| FrontierMath (Feb 2025 set) | 28.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- GPT-5.4 mini: 51.4 (#62)
- Llama 3-70B: 20.8 (#272)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| GPQA Diamond | 86.9% | 40.6% |
| LMArena Expert | 1434 | 1149 |
| SimpleQA Verified | 29.4% | — |
| Vectara Hallucination Rate | 5.5% | — |
| MMLU | — | 79.3% |

## Multimodal

- GPT-5.4 mini: 39.0 (#51)
- Llama 3-70B: —

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| LMArena Vision | 1245 | — |

## Multilingual

- GPT-5.4 mini: 52.1 (#94)
- Llama 3-70B: 33.8 (#246)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| LMArena Non-English | 1405 | 1142 |
| LMArena Chinese | 1446 | 1114 |
| LMArena French | 1440 | 1231 |
| LMArena German | 1407 | 1169 |
| LMArena Japanese | 1375 | 1017 |
| LMArena Korean | 1369 | 1017 |
| LMArena Russian | 1417 | 1159 |
| LMArena Spanish | 1403 | 1241 |

## Instruction Following

- GPT-5.4 mini: 74.3 (#99)
- Llama 3-70B: 62.8 (#233)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| LMArena Instruction Following | 1405 | 1194 |

## Long Context

- GPT-5.4 mini: 42.9 (#107)
- Llama 3-70B: 35.6 (#235)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| LMArena Longer Query | 1407 | 1174 |

## Writing & Preference

- GPT-5.4 mini: 64.2 (#56)
- Llama 3-70B: 43.0 (#225)

| Benchmark | GPT-5.4 mini | Llama 3-70B |
|---|---|---|
| LMArena Text | 1411 | 1221 |
| LMArena Creative Writing | 1370 | 1210 |
| LMArena Multi-Turn | 1429 | 1223 |
| EQ-Bench Creative Writing | 1665 | — |

## FAQ

### Is GPT-5.4 mini better than Llama 3-70B?

GPT-5.4 mini is the stronger model overall, scoring 44.8 to 28.5 on the Noometry Index.

### Is GPT-5.4 mini or Llama 3-70B better for coding?

GPT-5.4 mini scores higher on coding benchmarks: 45.0 versus 35.3 in the Noometry coding category.

### How many benchmarks do GPT-5.4 mini and Llama 3-70B share?

23 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Llama 3-70B has 31.