智商指数
87.8 #31
AAQI 62.0
输出速度
62 t/s
ITL 16.1 ms/tok
混合价格 (3:1)
$6.5 /1M
输入 $2/M · 输出 $10/M
上下文
262K
≈ 655 页 A4 · 最大输出 32,768
Quality Domain Radar
6-Axis Frontier Evaluation vs Category Average & Flagship
Simba 3.2
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
- Strong reasoning-per-dollar vs. flagship tiers
- Reliable multi-step tool calling (τ²-class)
- 262K context at mid-tier pricing
需要注意的权衡
- Smaller ecosystem — fewer SDKs and guides
- 62 tok/s mid-pack speed
- Text-only
杀手级用例
Autonomous research agents
Multi-step search-and-synthesize workflows.
Tool orchestration
Function calling over large API surfaces.
Analyst copilots
Reasoned synthesis across 262K-token briefs.
开发者快速上手与 API
请先设置
OPENROUTER_API_KEY — 获取 API 密钥curl https://openrouter.ai/api/v1/chat/completions \
-H "Authorization: Bearer $OPENROUTER_API_KEY" \
-H "content-type: application/json" \
-d '{
"model": "simba/simba-3-2",
"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://openrouter.ai/api/v1",
api_key=os.environ["OPENROUTER_API_KEY"],
)
resp = client.chat.completions.create(
model="simba/simba-3-2",
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://openrouter.ai/api/v1/chat/completions", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.OPENROUTER_API_KEY}`,
"content-type": "application/json",
},
body: JSON.stringify({
model: "simba/simba-3-2",
messages: [{ role: "user", content: "Refactor this function for O(n) time and explain the change." }],
}),
});
console.log((await res.json()).choices[0].message.content);OpenRouter-compatible endpoint — one API for 300+ models.
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
Simba 3.2 — reasoning frontier model
- Adjustable reasoning budget with verifier-guided decoding
- 262K context for multi-document tasks
- Tool-use fine-tune targeting multi-step agents
- 32K max output for long structured answers
发布日期
2026年10月5日
架构
Simba 3.2 — reasoning frontier model
参数量
Not disclosed
开源协议
Proprietary
上下文窗口
262,000 tokens
最大输出
32,768 tokens
模态
text
缓存价格
读 $0.2 · 写 $2.5 /1M
推理
思考力度档位
输入价格
$2/M
输出价格
$10/M
输出速度
62 t/s
首Token延迟
—
数据更新
2026年10月22日
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{llmpodium2026simba32,
title = {Benchmark Analysis and Intelligence Evaluation of Simba 3.2},
author = {LLMPodium Research},
year = {2026},
url = {https://llmpodium.com/models/simba-3-2}
}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 Independent 模型
替代模型推荐
基于能力、定价和基准测试得分与 Simba 3.2 最接近的替代模型。