Qwen3-4B
| Released | 2025-04-29 |
|---|---|
| Price per 1M | Not listed |
| Overall score | 33.4 (early, few results) |
| Provider | Alibaba |
| Weights | Open weights |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 33.4 | 2 | |
| Overall: Coding | 38.4 | 1 | |
| Overall: Reasoning | 38.4 | 1 | |
| Overall: Agentic work | 38.4 | 1 | |
| Reasoning | 38.4 | 1 | |
| General | 26.0 | 1 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 33.4 ±5.9 | 2 |
| Non-reasoning | 35.0 ±7.1 | 1 |
Evidence
Reasoning 38.4 ±4.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| GPQA Diamond | 52.3% | Epoch AI | 2026-08-28 |
General 26.0 ±5.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| EQ-Bench Longform Creative Writing | 17.0% | EQ-Bench | 2026-11-11 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| GPQA Diamond | 43.6% | Epoch AI | A stronger configuration result was used |
Names matched to this model
qwen3-4b