# Gemma 3 4B vs Mercury 2

> Mercury 2 is the stronger model overall, scoring 39.1 to 28.1 on the Noometry Index. Gemma 3 4B costs 7.5× 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/gemma-3-4b-vs-mercury-2
- Last updated: 2026-10-11
- Shared benchmarks: 11

## Summary

- They share 11 benchmarks with published results for both. Gemma 3 4B scores higher in 1 category and Mercury 2 in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Mercury 2 leads 36.2 to 11.8.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 6.4% for Gemma 3 4B and 12.3% for Mercury 2.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.25 / $0.75 for Mercury 2.
- Gemma 3 4B accepts more context: 131K tokens versus 128K.
- Gemma 3 4B has downloadable open weights; the other is API-only.

## Snapshot

| | Gemma 3 4B | Mercury 2 |
|---|---|---|
| Provider | Google | Inception |
| Noometry Index | 28.1 | 39.1 |
| Rank | 326 | 175 |
| Context | 131K | 128K |
| Input $/M | $0.04 | $0.25 |
| Output $/M | $0.08 | $0.75 |
| Weights | Open | Proprietary |

## Coding

- Gemma 3 4B: 35.9 (#215)
- Mercury 2: 33.5 (#255)

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| LMArena Coding | 1230 | 1391 |
| LMArena WebDev | — | 1171 |
| SciCode | — | 38.7% |
| WeirdML | — | 43.2% |
| ALE-Bench | — | 785.58 |

## Agentic & Tool Use

- Gemma 3 4B: 20.9 (#142)
- Mercury 2: —

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |

## Reasoning

- Gemma 3 4B: 13.2 (#335)
- Mercury 2: 23.8 (#170)

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| LMArena Hard Prompts | 1253 | 1362 |
| Kagi LLM Benchmark | 25.2% | — |
| CritPt | — | 0.8% |
| Chess Puzzles | 0% | — |
| DTBench | 50.9% | — |
| LMCA | 2.8% | — |
| Epoch Capabilities Index | 116.02 | — |

## Math

- Gemma 3 4B: 16.8 (#292)
- Mercury 2: —

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | — |
| LMArena Math | 1239 | — |

## Knowledge

- Gemma 3 4B: 11.8 (#299)
- Mercury 2: 36.2 (#172)

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| Vectara Hallucination Rate | 6.4% | 12.3% |
| LMArena Expert | 1223 | 1358 |
| GPQA Diamond | 23.2% | — |

## Multilingual

- Gemma 3 4B: 42.5 (#194)
- Mercury 2: 46.6 (#157)

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| LMArena Non-English | 1273 | 1331 |
| LMArena Russian | 1294 | 1304 |
| LMArena Chinese | — | 1417 |
| LMArena German | 1281 | — |

## Instruction Following

- Gemma 3 4B: 65.2 (#225)
- Mercury 2: 70.2 (#165)

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| LMArena Instruction Following | 1239 | 1329 |

## Long Context

- Gemma 3 4B: 38.7 (#194)
- Mercury 2: 40.5 (#154)

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| LMArena Longer Query | 1273 | 1330 |

## Writing & Preference

- Gemma 3 4B: 42.0 (#239)
- Mercury 2: 53.8 (#155)

| Benchmark | Gemma 3 4B | Mercury 2 |
|---|---|---|
| LMArena Text | 1291 | 1355 |
| LMArena Creative Writing | 1271 | 1289 |
| LMArena Multi-Turn | 1255 | 1358 |
| EQ-Bench Creative Writing | 1068 | — |

## FAQ

### Is Gemma 3 4B better than Mercury 2?

Mercury 2 is the stronger model overall, scoring 39.1 to 28.1 on the Noometry Index. Gemma 3 4B costs 7.5× less per token, which makes it the better buy when Mercury 2's lead doesn't matter for your workload.

### Which is cheaper, Gemma 3 4B or Mercury 2?

Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Mercury 2 lists at $0.25 and $0.75.

### Is Gemma 3 4B or Mercury 2 better for coding?

Gemma 3 4B scores higher on coding benchmarks: 35.9 versus 33.5 in the Noometry coding category.

### Which has the bigger context window?

Gemma 3 4B does, with 131K tokens against 128K.

### How many benchmarks do Gemma 3 4B and Mercury 2 share?

11 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Mercury 2 has 17.