Qwen
Qwen
Language model · China ·Apr 2025

Qwen3-14B

Language Open (Restricted)
14.8B
Parameters
46
Benchmarks

About Qwen3-14B

Qwen3-14B is an AI model developed by Qwen, in the language category, released in 2025, made available as an open-weights model with 14.8B params.

On this page you'll find Qwen3-14B's full specifications, including 46 benchmark results. Review provider pricing and benchmark scores below, or compare Qwen3-14B head-to-head with other language models.

Links

Run Qwen3-14B locally — what to buy

Share this build ↗

Sized to what Qwen3-14B actually needs (~8k context at Q4), not to the biggest GPU: Good = cheapest that runs it, Better = best value, Best = most headroom. Speed is a hardware estimate; anything past ~120 tok/s is shown as instant.

Software support

✓ measured · · compatible · — not supported. Informational only — speed is hardware-based.

Community

Would you run Qwen3-14B again?
No votes yet — be the first.
Did it run? — community reports by hardware

No reports yet. Own this model on some hardware? Be the first to confirm it runs.

Discussion

No comments yet. Start the discussion.

Qwen3-14B benchmark scores
Benchmark Category Score Variant
Codeforces Elo code 1,766 thinking-rating
Codeforces Elo code 1,200 non-thinking-rating
LiveCodeBench v5 code 63.5% thinking-pass@1
LiveCodeBench v5 code 29% non-thinking-pass@1
AlignBench v1.1 general 8.56 thinking-v1-1
AlignBench v1.1 general 8.52 non-thinking-v1-1
ArenaHard general 91.7% thinking
ArenaHard general 86.3% non-thinking
C-Eval general 86.2% thinking
C-Eval general 81% non-thinking
Creative Writing v3 general 80.3 thinking
Creative Writing v3 general 73.1 non-thinking
IFEval general 85.4% thinking-strict-prompt
IFEval general 84.8% non-thinking-strict-prompt
LiveBench general 71.3 thinking-2024-11-25
LiveBench general 59.6 non-thinking-2024-11-25
MMLU-Redux general 88.6% thinking
MMLU-Redux general 82% non-thinking
MMMLU general 77.9% thinking-14lang
MMMLU general 72.6% non-thinking-14lang
Multi-IF general 74.8 thinking
Multi-IF general 72.9 non-thinking
WritingBench general 7.8 thinking
WritingBench general 7.24 non-thinking
INCLUDE knowledge 71.7 thinking
INCLUDE knowledge 67.8 non-thinking
AIME 2024 math 79.3% thinking
AIME 2024 math 31.7% non-thinking
AIME 2025 math 70.4% thinking
AIME 2025 math 23.3% non-thinking
MATH-500 math 96.8% thinking
MATH-500 math 90% non-thinking
MT-AIME 2024 math 73.3 thinking-mt-2024
MT-AIME 2024 math 23.2 non-thinking-mt-2024
PolyMath-en math 45.8 thinking
PolyMath-en math 22 non-thinking
AutoLogi reasoning 89.2 thinking
AutoLogi reasoning 82 non-thinking
GPQA Diamond reasoning 64% thinking
GPQA Diamond reasoning 54.8% non-thinking
MLogiQA reasoning 71.1 thinking
MLogiQA reasoning 58.9 non-thinking
ZebraLogic reasoning 88.5% thinking
ZebraLogic reasoning 33% non-thinking
BFCL v3 tool-use 70.4% thinking
BFCL v3 tool-use 61.5% non-thinking

More models like Qwen3-14B

Frequently asked questions

What is Qwen3-14B?

Qwen3-14B is an AI model developed by Qwen, in the language category, released in 2025. It is tracked on GenAIList with its specifications, benchmark scores and provider pricing.

Who created Qwen3-14B?

Qwen3-14B was developed by Qwen and released in 2025.

Is Qwen3-14B open source or proprietary?

Qwen3-14B ships as an open-weights model: you can download and self-host the weights, though the licence may place some restrictions on use.

How much does Qwen3-14B cost?

Pricing for Qwen3-14B depends on the provider. See the providers table on this page for the latest API rates.

How does Qwen3-14B perform on benchmarks?

Qwen3-14B is benchmarked across 46 evaluations on GenAIList, including Codeforces Elo (1,766). See the full benchmark table below and compare it with other models.

No videos yet. Know a good one? Add it below.

Know a great video about Qwen3-14B?