# GPT-5.3 Codex vs Qwen3 32B

> GPT-5.3 Codex is the stronger model overall, scoring 45.5 to 38.8 on the Noometry Index. Qwen3 32B costs 3.9× less per token, which makes it the better buy when GPT-5.3 Codex's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-5-3-codex-vs-qwen3-32b
- Last updated: 2026-10-10
- Shared benchmarks: 1

## Summary

- They share 1 benchmark with published results for both. GPT-5.3 Codex scores higher in 2 categories and Qwen3 32B in 0 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GPT-5.3 Codex leads 47.8 to 32.3.
- Qwen3 32B is cheaper at $0.70 / $2.80 per million input/output tokens, against $1.75 / $14 for GPT-5.3 Codex.
- GPT-5.3 Codex accepts more context: 400K tokens versus 131K.
- Qwen3 32B has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 45.5 | 38.8 |
| Rank | 61 | 166 |
| Context | 400K | 131K |
| Input $/M | $1.75 | $0.70 |
| Output $/M | $14 | $2.80 |
| Weights | Proprietary | Open |

## Coding

- GPT-5.3 Codex: 48.5 (#51)
- Qwen3 32B: 37.3 (#184)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| SWE-bench Verified | 74.8% | — |
| Aider Polyglot | — | 40% |
| LMArena WebDev | 1408 | — |
| SciCode | — | 35.4% |
| WeirdML | 79.3% | — |
| LMArena Coding | — | 1358 |
| ALE-Bench | 1,655 | — |

## Agentic & Tool Use

- GPT-5.3 Codex: 47.8 (#9)
- Qwen3 32B: 32.3 (#58)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| Terminal-Bench | 78.4% | — |
| Berkeley Function Calling Leaderboard | — | 48.7% |
| METR Time Horizons | 74.5% | — |
| Vending-Bench 2 | 5,940 | — |

## Reasoning

- GPT-5.3 Codex: —
- Qwen3 32B: 19.6 (#233)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| Epoch Capabilities Index | 156.84 | 138.51 |
| Kagi LLM Benchmark | — | 54.9% |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 5% |
| LMArena Hard Prompts | — | 1335 |
| DTBench | — | 67.5% |
| LMCA | — | 17.3% |

## Math

- GPT-5.3 Codex: —
- Qwen3 32B: 37.9 (#89)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 66.9% |
| LMArena Math | — | 1399 |

## Knowledge

- GPT-5.3 Codex: —
- Qwen3 32B: 40.0 (#112)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| GPQA Diamond | — | 65.7% |
| Vectara Hallucination Rate | — | 5.9% |
| LMArena Expert | — | 1362 |

## Multilingual

- GPT-5.3 Codex: —
- Qwen3 32B: 45.8 (#162)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| LMArena Non-English | — | 1317 |
| LMArena Chinese | — | 1357 |
| LMArena German | — | 1341 |
| LMArena Russian | — | 1311 |

## Instruction Following

- GPT-5.3 Codex: —
- Qwen3 32B: 69.1 (#174)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | — | 1305 |

## Long Context

- GPT-5.3 Codex: —
- Qwen3 32B: 43.7 (#83)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| Fiction.LiveBench | — | 74.2% |
| LMArena Longer Query | — | 1327 |

## Writing & Preference

- GPT-5.3 Codex: —
- Qwen3 32B: 53.1 (#158)

| Benchmark | GPT-5.3 Codex | Qwen3 32B |
|---|---|---|
| LMArena Text | — | 1340 |
| LMArena Creative Writing | — | 1297 |
| LMArena Multi-Turn | — | 1331 |

## FAQ

### Is GPT-5.3 Codex better than Qwen3 32B?

GPT-5.3 Codex is the stronger model overall, scoring 45.5 to 38.8 on the Noometry Index. Qwen3 32B costs 3.9× less per token, which makes it the better buy when GPT-5.3 Codex's lead doesn't matter for your workload.

### Which is cheaper, GPT-5.3 Codex or Qwen3 32B?

Qwen3 32B is cheaper. It lists at $0.70 per million input tokens and $2.80 per million output tokens; GPT-5.3 Codex lists at $1.75 and $14.

### Is GPT-5.3 Codex or Qwen3 32B better for coding?

GPT-5.3 Codex scores higher on coding benchmarks: 48.5 versus 37.3 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.3 Codex does, with 400K tokens against 131K.

### How many benchmarks do GPT-5.3 Codex and Qwen3 32B share?

1 benchmark has published results for both models. GPT-5.3 Codex has 8 scored results on Noometry and Qwen3 32B has 26.