Model comparison
GPT-5 Nano vs Qwen3 14B
Qwen3 14B is the stronger model overall, scoring 35.0 to 33.4 on the Noometry Index. GPT-5 Nano costs 4.5× less per token, which makes it the better buy when Qwen3 14B's lead doesn't matter for your workload.
Last verified . 10 shared benchmarks.
Alibaba (Qwen)Summary
- They share 10 benchmarks with published results for both. GPT-5 Nano scores higher in 0 categories and Qwen3 14B in 6 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3 14B leads 37.0 to 29.1.
- The biggest single-benchmark swing is Chess Puzzles: 27% for GPT-5 Nano and 4% for Qwen3 14B.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $0.35 / $1.40 for Qwen3 14B.
- GPT-5 Nano accepts more context: 400K tokens versus 131K.
- Qwen3 14B has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Nano | Qwen3 14B | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 33.4 | 35.0 |
| Released | 2025-08-07 | 2025-04 |
| Weights | Proprietary | Open |
| Context window | 400K | 131K |
| Max output | 128K | 8K |
| Input $ / M tokens | $0.05 | $0.35 |
| Output $ / M tokens | $0.40 | $1.40 |
| Results tracked | 48 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 14B leads
GPT-5 Nano: 33.3 (#248), Qwen3 14B: 36.7 (#191)
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| SWE-bench Verified (bash only) | 34.8% | — |
| SciCode | — | 31.6% |
| WeirdML | 38.1% | — |
| LMArena Coding | 1351 | — |
| ALE-Bench | 718.67 | — |
Agentic & Tool Use Qwen3 14B leads
GPT-5 Nano: 25.6 (#99), Qwen3 14B: 29.4 (#78)
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 51.5% | 41% |
| Terminal-Bench | 21.8% | — |
Reasoning Qwen3 14B leads
GPT-5 Nano: 16.4 (#291), Qwen3 14B: 18.1 (#272)
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| Kagi LLM Benchmark | 62.2% | 49.1% |
| Chess Puzzles | 27% | 4% |
| DTBench | 62.7% | 64% |
| LMCA | 7.9% | 18.2% |
| Epoch Capabilities Index | 139.39 | 138.23 |
| ARC-AGI-2 | 2.6% | — |
| ARC-AGI-1 | 20.7% | — |
| CritPt | — | 0% |
| LMArena Hard Prompts | 1329 | — |
| Mystery Game Puzzles | 9% | — |
| ForecastBench | 59.1 | — |
Math Qwen3 14B leads
GPT-5 Nano: 29.1 (#221), Qwen3 14B: 37.0 (#102)
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 81.1% | 66.4% |
| FrontierMath (Tiers 1-3) | 20% | — |
| FrontierMath Tier 4 | 2.4% | — |
| ProofBench | 12% | — |
| Omni-MATH | 54.6% | — |
| LMArena Math | 1318 | — |
| MATH Level 5 | 95.2% | — |
| FrontierMath (Feb 2025 set) | 8.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge Qwen3 14B leads
GPT-5 Nano: 35.8 (#166), Qwen3 14B: 39.2 (#124)
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| GPQA Diamond | 69.4% | 63.8% |
| Vectara Hallucination Rate | 10.5% | 5.4% |
| SimpleQA Verified | 11.7% | — |
| MMLU-Pro | 77.8% | — |
| GPQA (HELM) | 67.9% | — |
| LMArena Expert | 1321 | — |
Multimodal Not comparable
GPT-5 Nano: 30.7 (#102), Qwen3 14B: —
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| LMArena Vision | 1159 | — |
| VPCT | 37.2% | — |
Multilingual Not comparable
GPT-5 Nano: 45.5 (#167), Qwen3 14B: —
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| LMArena Non-English | 1313 | — |
| LMArena Chinese | 1355 | — |
| LMArena German | 1329 | — |
| LMArena Korean | 1269 | — |
| LMArena Russian | 1296 | — |
| LMArena Spanish | 1360 | — |
Instruction Following Not comparable
GPT-5 Nano: 75.2 (#72), Qwen3 14B: —
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| IFEval | 93.2% | — |
| LMArena Instruction Following | 1306 | — |
Long Context Qwen3 14B leads
GPT-5 Nano: 31.3 (#275), Qwen3 14B: 38.0 (#201)
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| Fiction.LiveBench | 44.4% | 62.5% |
| LMArena Longer Query | 1313 | — |
Writing & Preference Not comparable
GPT-5 Nano: 39.4 (#244), Qwen3 14B: —
| Benchmark | GPT-5 Nano | Qwen3 14B |
|---|---|---|
| LMArena Text | 1320 | — |
| LMArena Creative Writing | 1250 | — |
| EQ-Bench Creative Writing | 705 | — |
| WildBench | 80.6% | — |
| LMArena Multi-Turn | 1311 | — |
Frequently asked questions
Is GPT-5 Nano better than Qwen3 14B?
Qwen3 14B is the stronger model overall, scoring 35.0 to 33.4 on the Noometry Index. GPT-5 Nano costs 4.5× less per token, which makes it the better buy when Qwen3 14B's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Nano or Qwen3 14B?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Qwen3 14B lists at $0.35 and $1.40.
Is GPT-5 Nano or Qwen3 14B better for coding?
Qwen3 14B scores higher on coding benchmarks: 36.7 versus 33.3 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 131K.
How many benchmarks do GPT-5 Nano and Qwen3 14B share?
10 benchmarks have published results for both models. GPT-5 Nano has 48 scored results on Noometry and Qwen3 14B has 12.