Qwen3.6 27B
Qwen3.6 27BRankingin thinking mode?
Alternatives?
CheaperDeepSeek V4.1 Flash 1017maxOpen+9%-61%$0.26Memory (4-bit): 17 GB → 458 GB+5
BetterGPT-6 Lunamax+9%-85%$0.068Function calling only with none · Chat Completions+3OpenQwen3.6 35B A3Bunstated-5%-59%$0.25 / $1.49Memory (4-bit): 17 GB → 22 GB
CheaperClaude Haiku 5.5xhigh+9%-85%$0.12Remove sampling parameters+5CheaperGPT-5.6 Lunamax+8%-67%$0.18DeepSeek V4.1 Flash 1017maxOpen+9%-61%$0.26Memory (4-bit): 17 GB → 458 GB+5Claude Haiku 5.5xhigh+9%-85%$0.12Remove sampling parameters+5GPT-6 Lunamax+9%-85%$0.068Function calling only with none · Chat Completions+3GPT-5.6 Lunamax+8%-67%$0.18
Gemini 3 Flash Previewhigh+5%-17%$0.50 / $3GPT-6 Lunamax+9%-85%$0.068Function calling only with none · Chat Completions+3Claude Haiku 5.5high+7%-85%$0.079Remove sampling parameters+5GPT-5.6 Lunamax+8%-67%$0.18
Grok 4.3high+2%+16%$0.21GPT-5.6 Sollow+5%+493%$0.26Qwen3.6 35B A3BunstatedOpen-5%-59%$0.25 / $1.49Memory (4-bit): 17 GB → 22 GBGemma 4 26B A4BminimalOpen-10%-81%$0.15 / $0.60Context: 262K → 256KGemma 4 31BminimalOpen-20%-88%$0.10 / $0.34Context: 262K → 256K+1Qwen3 30B A3B 2507 (Reasoning)unstatedOpen-23%——Memory (4-bit): 17 GB → 18 GB+1Qwen 3.5 35B A3BnoneOpen-23%-49%$0.25 / $2Memory (4-bit): 17 GB → 22 GBQwen3.8 Max (0902)xhigh+10%—$2 / $6Qwen3.7 Maxmax+5%—$1.48 / $4.43No image inputQwen3.7 Plusunstated+2%—$0.32 / $1.28Qwen3.6 Plusunstated+2%—$0.50 / $3Qwen3.6 Max Previewunstatedsame—$1.30 / $7.80No image input+1Qwen3.5 397B A17BunstatedOpen-2%—$0.60 / $3.60Memory (4-bit): 17 GB → 242 GBQwen3.7 Flashunstated-5%—$0.03 / $0.13Qwen3.6 35B A3BunstatedOpen-5%—$0.25 / $1.49Memory (4-bit): 17 GB → 22 GBAll Alibaba (Qwen) models →Run it yourself
estimateBenchmarks
all categoriesFAQ
How much does Qwen3.6 27B cost?›
$0.60 per 1M input tokens and $3.60 per 1M output tokens; cheapest via Reka at $0.20 / $2. A typical benchmark task costs about $0.62 at default effort.
Is Qwen3.6 27B free to use?›
There is no free API; the cheapest paid price is $0.20 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Qwen3.6 27B, and what is it best at?›
#110 of 336 in Chat & Knowledge, #77 of 134 in Coding, #76 of 135 in Agents & Tool Use, #60 of 245 in Writing & Language, #54 of 125 in Vision & Documents. Strongest at: Tool calling, Long context, Instructions.
Is Qwen3.6 27B better than Claude Opus 4.6?›
On OTIS Mock AIME 2024 2025, Qwen3.6 27B scores 91.1 vs 91.1 — about the same.
How can I use Qwen3.6 27B?›
Through the API of Alibaba, Chutes, DeepInfra, Novita, Phala and 3 more.
What is the context window of Qwen3.6 27B?›
262K tokens, with up to 66K output tokens.
When was Qwen3.6 27B released? Is there a newer version?›
Released 2026-04-22. It has been superseded by Qwen3.8 27B.
Is Qwen3.6 27B open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Qwen3.6 27B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Qwen3.6 27B read images?›
It accepts text and images, video as input.
What is Qwen3.6 27B?›
Qwen3.6 27B is an open-weights language model from Alibaba (Qwen), released Apr 2026, the successor to Qwen 3.5 27B.
What are the best alternatives to Qwen3.6 27B?›
DeepSeek V4.1 Flash 1017 (cheaper), GPT-6 Luna (better), Qwen3.6 35B A3B (open weights).