# Nova 2 Lite vs Qwen2-72B

> Nova 2 Lite is the stronger model overall, scoring 39.7 to 30.0 on the Noometry Index.

- Canonical page: https://noometry.com/compare/nova-2-lite-vs-qwen2-72b
- Last updated: 2026-10-11
- Shared benchmarks: 17

## Summary

- They share 17 benchmarks with published results for both. Nova 2 Lite scores higher in 9 categories and Qwen2-72B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Nova 2 Lite leads 43.0 to 21.2.
- Qwen2-72B has downloadable open weights; the other is API-only.

## Snapshot

| | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| Provider | Amazon | Alibaba (Qwen) |
| Noometry Index | 39.7 | 30.0 |
| Rank | 161 | 300 |
| Context | 1M | — |
| Input $/M | $0.30 | — |
| Output $/M | $2.50 | — |
| Weights | Proprietary | Open |

## Coding

- Nova 2 Lite: 40.7 (#134)
- Qwen2-72B: 29.1 (#310)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Coding | 1385 | 1196 |
| WeirdML | — | 11.3% |
| BigCodeBench Instruct | — | 38.5% |
| BigCodeBench Complete | — | 54% |

## Agentic & Tool Use

- Nova 2 Lite: 24.1 (#121)
- Qwen2-72B: 17.0 (#146)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 27.1% | — |
| TheAgentCompany | — | 1.1% |
| METR Time Horizons | — | 29.9% |

## Reasoning

- Nova 2 Lite: 27.5 (#118)
- Qwen2-72B: 23.2 (#181)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1364 | 1191 |
| Epoch Capabilities Index | — | 125.28 |

## Math

- Nova 2 Lite: 37.5 (#156)
- Qwen2-72B: 30.2 (#236)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Math | 1359 | 1235 |
| MATH Level 5 | — | 39.1% |

## Knowledge

- Nova 2 Lite: 43.0 (#94)
- Qwen2-72B: 21.2 (#275)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Expert | 1358 | 1171 |
| GPQA Diamond | — | 40.8% |
| Vectara Hallucination Rate | 5.1% | — |
| MMLU | — | 82.4% |

## Multilingual

- Nova 2 Lite: 47.1 (#153)
- Qwen2-72B: 35.9 (#244)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1337 | 1176 |
| LMArena Chinese | 1364 | 1240 |
| LMArena French | 1381 | 1170 |
| LMArena German | 1343 | 1151 |
| LMArena Japanese | 1271 | 1111 |
| LMArena Korean | 1284 | 1083 |
| LMArena Russian | 1343 | 1169 |
| LMArena Spanish | 1373 | 1169 |

## Instruction Following

- Nova 2 Lite: 70.5 (#161)
- Qwen2-72B: 61.7 (#241)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1335 | 1181 |

## Long Context

- Nova 2 Lite: 40.6 (#150)
- Qwen2-72B: 36.1 (#235)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1335 | 1192 |

## Writing & Preference

- Nova 2 Lite: 53.9 (#154)
- Qwen2-72B: 40.8 (#241)

| Benchmark | Nova 2 Lite | Qwen2-72B |
|---|---|---|
| LMArena Text | 1362 | 1203 |
| LMArena Creative Writing | 1291 | 1181 |
| LMArena Multi-Turn | 1338 | 1196 |

## FAQ

### Is Nova 2 Lite better than Qwen2-72B?

Nova 2 Lite is the stronger model overall, scoring 39.7 to 30.0 on the Noometry Index.

### Is Nova 2 Lite or Qwen2-72B better for coding?

Nova 2 Lite scores higher on coding benchmarks: 40.7 versus 29.1 in the Noometry coding category.

### How many benchmarks do Nova 2 Lite and Qwen2-72B share?

17 benchmarks have published results for both models. Nova 2 Lite has 19 scored results on Noometry and Qwen2-72B has 26.