# Nova Premier 1.0 vs Qwen3-4B

> Nova Premier 1.0 is the stronger model overall, scoring 37.7 to 31.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/nova-premier-1-0-vs-qwen3-4b
- Last updated: 2026-10-10
- Shared benchmarks: 0

## Summary

- The widest gap is in reasoning, where Nova Premier 1.0 leads 22.3 to 18.5.
- Qwen3-4B has downloadable open weights; the other is API-only.

## Snapshot

| | Nova Premier 1.0 | Qwen3-4B |
|---|---|---|
| Provider | Amazon | Alibaba (Qwen) |
| Noometry Index | 37.7 | 31.5 |
| Rank | 185 | 261 |
| Context | 1M | — |
| Input $/M | $2.50 | — |
| Output $/M | $12.50 | — |
| Weights | Proprietary | Open |

## Agentic & Tool Use

- Nova Premier 1.0: —
- Qwen3-4B: 27.3 (#94)

| Benchmark | Nova Premier 1.0 | Qwen3-4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 35.7% |

## Reasoning

- Nova Premier 1.0: 22.3 (#182)
- Qwen3-4B: 18.5 (#264)

| Benchmark | Nova Premier 1.0 | Qwen3-4B |
|---|---|---|
| Kagi LLM Benchmark | 44.8% | — |
| Chess Puzzles | — | 4% |

## Math

- Nova Premier 1.0: 31.5 (#194)
- Qwen3-4B: 28.5 (#228)

| Benchmark | Nova Premier 1.0 | Qwen3-4B |
|---|---|---|
| MathArena Final-Answer Competitions | — | 38.5% |
| OTIS Mock AIME 2024-2025 | — | 52.2% |
| Omni-MATH | 35% | — |

## Knowledge

- Nova Premier 1.0: 35.4 (#176)
- Qwen3-4B: 33.0 (#198)

| Benchmark | Nova Premier 1.0 | Qwen3-4B |
|---|---|---|
| GPQA Diamond | — | 52.3% |
| MMLU-Pro | 72.6% | — |
| Vectara Hallucination Rate | — | 5.7% |
| GPQA (HELM) | 51.8% | — |

## Instruction Following

- Nova Premier 1.0: 66.9 (#201)
- Qwen3-4B: —

| Benchmark | Nova Premier 1.0 | Qwen3-4B |
|---|---|---|
| IFEval | 80.3% | — |

## Writing & Preference

- Nova Premier 1.0: 51.0 (#174)
- Qwen3-4B: —

| Benchmark | Nova Premier 1.0 | Qwen3-4B |
|---|---|---|
| WildBench | 78.8% | — |

## FAQ

### Is Nova Premier 1.0 better than Qwen3-4B?

Nova Premier 1.0 is the stronger model overall, scoring 37.7 to 31.5 on the Noometry Index.

### How many benchmarks do Nova Premier 1.0 and Qwen3-4B share?

0 benchmarks have published results for both models. Nova Premier 1.0 has 6 scored results on Noometry and Qwen3-4B has 6.