Qwen 3.5 9B
Qwen 3.5 9BAlternatives?
CheaperClaude Haiku 5.5xhighwins 100%-42%$0.12Remove sampling parameters+5
BetterGPT-6 Lunalowwins 91%-98%$0.005Function calling only with none · Chat Completions+3IOpenK2 Horizon 7Bthinkingwins 98%+468%$0.30 / $1.20No image input
CheaperMiMo v2.6 ProthinkingOpenwins 100%-37%$0.13Memory (4-bit): 4.8 GB → 615 GB+4CheaperMiMo v2.6 FlashthinkingOpenwins 100%-71%$0.062Memory (4-bit): 4.8 GB → 186 GB+4Claude Haiku 5.5xhighwins 100%-42%$0.12Remove sampling parameters+5MiMo v2.6 ProthinkingOpenwins 100%-37%$0.13Memory (4-bit): 4.8 GB → 615 GB+4MiMo v2.6 FlashthinkingOpenwins 100%-71%$0.062Memory (4-bit): 4.8 GB → 186 GB+4GPT-5.6 Lunamaxwins 100%-16%$0.18GPT-6 Lunamaxwins 100%-68%$0.068Function calling only with none · Chat Completions+3GPT-6 Lunalowwins 91%-98%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalowwins 91%-95%$0.01MiMo v2.5thinkingOpenwins 92%-91%$0.019Memory (4-bit): 4.8 GB → 186 GBClaude Haiku 5.5lowwins 99%-88%$0.024Remove sampling parameters+5GPT-5 Minihighwins 62%-75%$0.053
Ling 3.0 TinyOpenwins 68%——No image inputOMiniCPM5 2BthinkingOpenwins 62%——Context: 262K → 131K+1
Granite 4.2 8BthinkingOpenwins 46%-89%$0.024Context: 262K → 128K+1Qwen3.8 Max (0902)thinkingwins 100%+2,444%$5.41Qwen3.8 FlashthinkingOpenwins 100%+75%$0.37Memory (4-bit): 4.8 GB → 108 GB+6Qwen3.8 2.4T A95BthinkingOpenwins 100%+914%$2.16Memory (4-bit): 4.8 GB → 1,468 GB+6Qwen3.8 27BmediumOpenwins 100%+433%$1.13Memory (4-bit): 4.8 GB → 17 GB+5Qwen3.7 Maxthinkingwins 96%+440%$1.15No image inputQwen3.7 Plusthinkingwins 96%same$0.22Qwen3.6 27BthinkingOpenwins 94%+192%$0.62Memory (4-bit): 4.8 GB → 17 GBQwen3.6 35B A3BthinkingOpenwins 89%+123%$0.48Memory (4-bit): 4.8 GB → 22 GBAll Alibaba (Qwen) models →Run it yourself
estimateBenchmarks
in Agents & Tool UseFAQ
How much does Qwen 3.5 9B cost?›
$0.08 per 1M input tokens and $0.13 per 1M output tokens (cheapest host: Darkbloom). A typical benchmark task costs about $0.21 at default effort.
Is Qwen 3.5 9B free to use?›
There is no free API; the cheapest paid price is $0.08 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Qwen 3.5 9B, and what is it best at?›
#144 of 334 in Chat & Knowledge, #63 of 132 in Coding, #109 of 133 in Agents & Tool Use, #65 of 245 in Writing & Language, #72 of 124 in Vision & Documents. Strongest at: Tool calling.
Is Qwen 3.5 9B better than Command A+?›
On AA-Briefcase 1.1 · Elo, Qwen 3.5 9B scores 340.7 vs 355.9 — lower.
How can I use Qwen 3.5 9B?›
Through the API of Darkbloom, DeepInfra, Featherless, Parasail, SiliconFlow and 2 more.
What is the context window of Qwen 3.5 9B?›
262K tokens, with up to 33K output tokens.
When was Qwen 3.5 9B released? Is there a newer version?›
Released 2026-03-02. It is the current version in its line.
Is Qwen 3.5 9B open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Qwen 3.5 9B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Qwen 3.5 9B read images?›
It accepts text and images, video as input.
What is Qwen 3.5 9B?›
Qwen 3.5 9B is an open-weights language model from Alibaba (Qwen), released Mar 2026, the successor to Qwen3 VL 8B Instruct.
What are the best alternatives to Qwen 3.5 9B?›
Claude Haiku 5.5 (cheaper), GPT-6 Luna (better), K2 Horizon 7B (open weights).