# Claude 2 vs Qwen3-4B

> Qwen3-4B is the stronger model overall, scoring 31.9 to 25.0 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-2-vs-qwen3-4b
- Last updated: 2026-10-10
- Shared benchmarks: 2

## Summary

- They share 2 benchmarks with published results for both. Claude 2 scores higher in 1 category and Qwen3-4B in 2 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3-4B leads 29.7 to 9.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 2.5% for Claude 2 and 52.2% for Qwen3-4B.
- Qwen3-4B has downloadable open weights; the other is API-only.

## Snapshot

| | Claude 2 | Qwen3-4B |
|---|---|---|
| Provider | Anthropic | Alibaba (Qwen) |
| Noometry Index | 25.0 | 31.9 |
| Rank | 346 | 264 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Proprietary | Open |

## Coding

- Claude 2: —
- Qwen3-4B: —

| Benchmark | Claude 2 | Qwen3-4B |
|---|---|---|
| HumanEval+ | 61.6% | — |

## Agentic & Tool Use

- Claude 2: —
- Qwen3-4B: 27.6 (#100)

| Benchmark | Claude 2 | Qwen3-4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 35.7% |

## Reasoning

- Claude 2: 21.7 (#216)
- Qwen3-4B: 19.2 (#268)

| Benchmark | Claude 2 | Qwen3-4B |
|---|---|---|
| Chess Puzzles | — | 4% |
| DTBench | 51.9% | — |
| Epoch Capabilities Index | 120.13 | — |

## Math

- Claude 2: 9.3 (#320)
- Qwen3-4B: 29.7 (#240)

| Benchmark | Claude 2 | Qwen3-4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.5% | 52.2% |
| MathArena Final-Answer Competitions | — | 38.5% |
| MATH Level 5 | 11.7% | — |

## Knowledge

- Claude 2: 16.9 (#287)
- Qwen3-4B: 33.0 (#208)

| Benchmark | Claude 2 | Qwen3-4B |
|---|---|---|
| GPQA Diamond | 34.7% | 52.3% |
| Vectara Hallucination Rate | — | 5.7% |
| MMLU | 78.5% | — |
| TriviaQA | 87.5% | — |

## FAQ

### Is Claude 2 better than Qwen3-4B?

Qwen3-4B is the stronger model overall, scoring 31.9 to 25.0 on the Noometry Index.

### How many benchmarks do Claude 2 and Qwen3-4B share?

2 benchmarks have published results for both models. Claude 2 has 8 scored results on Noometry and Qwen3-4B has 6.