智商指数
88.2 #28
AAQI 88.0
输出速度
55 t/s
ITL 18.2 ms/tok
混合价格 (3:1)
$0.02 /1M
输入 $0.02/M · 输出 $0.02/M
上下文
64K
≈ 160 页 A4 · 最大输出 4,096
Quality Domain Radar
6-Axis Frontier Evaluation vs Category Average & Flagship
FLUX.2 Klein 9B
Claude Mythos Preview (#1)
Average (68)
缺少直接基准测量时,雷达图各轴由类别得分近似得出 — 详见 评分方法。
优势与取舍
优势所在
- Best quality-per-GB in the FLUX.2 family
- Runs on RTX 3060-class hardware
- Open weights — fine-tune and self-host freely
需要注意的权衡
- Still trails Dev/Pro on typography and faces
- Half the speed of the 4B klein tier
- 2K output ceiling
杀手级用例
Hobbyist & indie pipelines
Pro-adjacent images on consumer cards.
Fine-tuned niche styles
Small enough to iterate LoRAs overnight.
Local production tools
Self-hosted rendering for studios and agencies.
开发者快速上手与 API
请先设置
HF_TOKEN — 获取 API 密钥curl https://router.huggingface.co/v1/chat/completions \
-H "Authorization: Bearer $HF_TOKEN" \
-H "content-type: application/json" \
-d '{
"model": "black-forest-labs/FLUX.2-klein-9b",
"messages": [{"role": "user", "content": "Refactor this function for O(n) time and explain the change."}]
}'import os
from openai import OpenAI
client = OpenAI(
base_url="https://router.huggingface.co/v1",
api_key=os.environ["HF_TOKEN"],
)
resp = client.chat.completions.create(
model="black-forest-labs/FLUX.2-klein-9b",
messages=[{"role": "user", "content": "Refactor this function for O(n) time and explain the change."}],
)
print(resp.choices[0].message.content)const res = await fetch("https://router.huggingface.co/v1/chat/completions", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.HF_TOKEN}`,
"content-type": "application/json",
},
body: JSON.stringify({
model: "black-forest-labs/FLUX.2-klein-9b",
messages: [{ role: "user", content: "Refactor this function for O(n) time and explain the change." }],
}),
});
console.log((await res.json()).choices[0].message.content);官方资源与链接
API 部署
对比托管商与延迟
Groq、Cerebras、Together、DeepInfra、Fireworks、Bedrock 与 Azure 的吞吐量、TTFT 和缓存定价。
GroqCerebrasTogether AIDeepInfraFireworks AIAWS BedrockAzure AI
各榜单得分
分类能力得分
基准测试分项
实时规格与架构
底层解析
FLUX.2 Klein 9B — consumer-GPU diffusion
- 9B distilled transformer — the klein quality tier
- Fits 8GB VRAM at Q4
- Near-Pro prompt adherence on standard scenes
- Full LoRA fine-tuning support
发布日期
2026年10月14日
架构
FLUX.2 Klein 9B — consumer-GPU diffusion
参数量
9B
开源协议
Open Weights
上下文窗口
64,000 tokens
最大输出
4,096 tokens
模态
text + image
缓存价格
读 $0.008 · 写 $0.025 /1M
推理
—
输入价格
$0.02/M
输出价格
$0.02/M
输出速度
55 t/s
首Token延迟
—
数据更新
2026年10月22日
硬件与本地推理
数据来源:Convly AI(CC BY 4.0) ↗VRAM(4-bit Q4)
6 GB
最低 GPU
RTX 3060 8GB / any 8GB GPU
每月账单预估
每月预估支出
$0.00
正在计算相对 GPT-4o 的节省额…
引用与嵌入
GitHub Markdown 徽章
Loading badge markdown...BibTeX
@misc{llmpodium2026flux2klein9b,
title = {Benchmark Analysis and Intelligence Evaluation of FLUX.2 Klein 9B},
author = {LLMPodium Research},
year = {2026},
url = {https://llmpodium.com/models/flux-2-klein-9b}
}嵌入评分徽章
将实时 Podium 评分徽章添加到 GitHub 或 Hugging Face README —— 链接回实时排名。
预览
Markdown
Loading badge snippet...HTML
Loading badge snippet...更多 Black Forest Labs 模型
替代模型推荐
基于能力、定价和基准测试得分与 FLUX.2 Klein 9B 最接近的替代模型。