Qwen
QwenCompare inon
Benchmarks of Qwen14
BenchmarkByScoreRank
No benchmark matches
—per task?
price not verifiedprice
—output speed?
—context
Ranking?
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-5 Minihighwins 62%—$0.053
OpenMiMo v2.6 Prounstatedwins 77%—$0.43 / $0.87Pass back reasoning_content+3BetterGPT-6 Lunamaxwins 64%—$0.068Function calling only with none · Chat Completions+3
BetterClaude Haiku 5.5highwins 66%—$0.079Remove sampling parameters+5BetterGPT-5.6 Lunaxhighwins 70%—$0.085ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-5 Minihighwins 62%—$0.053GPT-6 Lunamaxwins 64%—$0.068Function calling only with none · Chat Completions+3Claude Haiku 5.5highwins 66%—$0.079Remove sampling parameters+5GPT-5.6 Lunaxhighwins 70%—$0.085GPT-5.6 Terraxhighwins 71%—$0.63Prompts with >272K input tokensModelScoreCostPer taskWhat changes
MiMo v2.6 ProunstatedOpenwins 77%—$0.43 / $0.87Pass back reasoning_content+3
GLM 5.3maxOpenwins 75%—$2.01Reasoning cannot be disabled+3
Kimi K3maxOpenwins 75%—$2.00Sampling parameters are fixed+3MiMo v2.5 ProunstatedOpenwins 74%—$0.43 / $0.87GLM 5.1unstatedOpenwins 74%—$1.40 / $4.40ModelScoreCostPer taskWhat changes
Qwen 3.7 Max Previewwins 74%—$2.50 / $7.50Qwen3.7 Plusunstatedwins 72%—$0.32 / $1.28Qwen3.6 Max Previewunstatedwins 71%—$1.30 / $7.80Include the word JSON · QwenCloudQwen3.8 27BunstatedOpenwins 71%—$0.42 / $2.55max_tokens range · QwenCloud+4Qwen3.6 Plusunstatedwins 71%—$0.50 / $3Qwen3.5 397B A17BunstatedOpenwins 70%—$0.60 / $3.60Qwen3 Max (Preview)unstatedwins 70%—$0.87 / $3.48Qwen3 235B A22B Thinking 2507unstatedOpenwins 69%—$0.23 / $2.30No strict JSONAll Alibaba (Qwen) models →Run it yourself
Fits onhardware needs not verified yet
Licencenot verified
Benchmarks
all categoriesChat & Knowledge
Arena Text · overall1,318#145
Arena Text · Chinese1,254.1#160
Arena Text · English1,292.5#168
Arena Text · French1,330.2#127
Arena Text · German1,253.9#133
Arena Text · instruction following1,261.7#159
Arena Text · Japanese1,205.1#122
Arena Text · Korean1,141.5#139
Arena Text · non-English1,262.9#159
Arena Text · Russian1,273.9#154
Arena Text · Spanish1,290.1#138
Arena Text · overall · October 22, 20261,281.9#156
Writing & Language
Arena Text · creative writing1,285#139
FAQ
How much does Qwen cost?›
There is no per-unit API price: price not verified.
How good is Qwen, and what is it best at?›
#168 of 205 in Chat & Knowledge, #155 of 197 in Writing & Language.
Is Qwen better than Llama 3.1 70B Instruct?›
On Arena Text · English, Qwen scores 1,292.5 vs 1,293.6 — about the same.
Is Qwen open source?›
Yes, the weights are open.
Can I run Qwen locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Qwen read images?›
No, it accepts text only.
What is Qwen?›
Qwen is an open-weights language model from Alibaba (Qwen).
What are the best alternatives to Qwen?›
GPT-5 Mini (better), MiMo v2.6 Pro (open weights).