# Llama-3.3-70B-Instruct vs Mercury 2

> Mercury 2 is the stronger model overall, scoring 38.8 to 30.3 on the Noometry Index. Llama-3.3-70B-Instruct costs 2.4× less per token, which makes it the better buy when Mercury 2's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/llama-3-3-70b-instruct-vs-mercury-2
- Last updated: 2026-10-10
- Shared benchmarks: 15

## Summary

- They share 15 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 1 category and Mercury 2 in 6 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in long context, where Mercury 2 leads 40.4 to 26.4.
- The biggest single-benchmark swing is WeirdML: 14.4% for Llama-3.3-70B-Instruct and 43.2% for Mercury 2.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $0.25 / $0.75 for Mercury 2.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.

## Snapshot

| | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| Provider | Meta | Inception |
| Noometry Index | 30.3 | 38.8 |
| Rank | 283 | 167 |
| Context | 128K | 128K |
| Input $/M | $0.10 | $0.25 |
| Output $/M | $0.32 | $0.75 |
| Weights | Open | Proprietary |

## Coding

- Llama-3.3-70B-Instruct: 30.6 (#284)
- Mercury 2: 33.2 (#250)

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| SciCode | 26% | 38.7% |
| WeirdML | 14.4% | 43.2% |
| LMArena Coding | 1268 | 1391 |
| LMArena WebDev | — | 1170 |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
| BigCodeBench Complete | 57.5% | — |
| ALE-Bench | — | 785.58 |

## Agentic & Tool Use

- Llama-3.3-70B-Instruct: 25.6 (#100)
- Mercury 2: —

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | — |
| BALROG | 23% | — |

## Reasoning

- Llama-3.3-70B-Instruct: 13.7 (#325)
- Mercury 2: 23.0 (#170)

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| CritPt | 0% | 0.8% |
| LMArena Hard Prompts | 1257 | 1362 |
| SimpleBench | 19.9% | — |
| LiveBench Reasoning | 50.8% | — |
| DTBench | 59.5% | — |
| LiveBench Data Analysis | 49.5% | — |
| LMCA | 17.5% | — |
| Epoch Capabilities Index | 127.32 | — |
| ForecastBench | 58.6 | — |
| LiveBench | 50.2% | — |

## Math

- Llama-3.3-70B-Instruct: 14.3 (#292)
- Mercury 2: —

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 5.1% | — |
| LiveBench Math | 42.2% | — |
| LMArena Math | 1268 | — |
| MATH Level 5 | 41.6% | — |

## Knowledge

- Llama-3.3-70B-Instruct: 30.5 (#220)
- Mercury 2: 35.6 (#170)

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| Vectara Hallucination Rate | 4.1% | 12.3% |
| LMArena Expert | 1225 | 1358 |
| GPQA Diamond | 47.4% | — |
| Confabulations | 22.8% | — |
| MMLU | 86.3% | — |

## Multilingual

- Llama-3.3-70B-Instruct: 40.1 (#215)
- Mercury 2: 46.8 (#153)

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| LMArena Non-English | 1236 | 1331 |
| LMArena Chinese | 1217 | 1416 |
| LMArena Russian | 1252 | 1303 |
| LMArena French | 1280 | — |
| LMArena German | 1251 | — |
| LMArena Japanese | 1149 | — |
| LMArena Korean | 1142 | — |
| LMArena Spanish | 1269 | — |

## Instruction Following

- Llama-3.3-70B-Instruct: 71.3 (#152)
- Mercury 2: 70.4 (#161)

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| LMArena Instruction Following | 1242 | 1328 |
| LiveBench Instruction Following | 82.7% | — |

## Long Context

- Llama-3.3-70B-Instruct: 26.4 (#290)
- Mercury 2: 40.4 (#149)

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| LMArena Longer Query | 1256 | 1329 |
| Fiction.LiveBench | 33.3% | — |

## Writing & Preference

- Llama-3.3-70B-Instruct: 47.8 (#203)
- Mercury 2: 54.0 (#151)

| Benchmark | Llama-3.3-70B-Instruct | Mercury 2 |
|---|---|---|
| LMArena Text | 1274 | 1355 |
| LMArena Creative Writing | 1250 | 1289 |
| LMArena Multi-Turn | 1279 | 1358 |
| LiveBench Language | 39.2% | — |

## FAQ

### Is Llama-3.3-70B-Instruct better than Mercury 2?

Mercury 2 is the stronger model overall, scoring 38.8 to 30.3 on the Noometry Index. Llama-3.3-70B-Instruct costs 2.4× less per token, which makes it the better buy when Mercury 2's lead doesn't matter for your workload.

### Which is cheaper, Llama-3.3-70B-Instruct or Mercury 2?

Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; Mercury 2 lists at $0.25 and $0.75.

### Is Llama-3.3-70B-Instruct or Mercury 2 better for coding?

Mercury 2 scores higher on coding benchmarks: 33.2 versus 30.6 in the Noometry coding category.

### Which has the bigger context window?

Both accept 128K tokens.

### How many benchmarks do Llama-3.3-70B-Instruct and Mercury 2 share?

15 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and Mercury 2 has 17.