Ministral 8B vs Ministral 3B
Benchmarks, API Token Pricing, Context Window & Speed Comparison
Related topics (12)
Mistral AI 的 Ministral 8B 与 Mistral AI 的 Ministral 3B 是当前备受关注的两款主力 AI 模型。Ministral 8B 提供 128K Token 上下文窗口，API 定价为每百万 Token 输入 $0.10 / 输出 $0.10；而 Ministral 3B 提供 128K Token 上下文窗口，定价为每百万 Token 输入 $0.04 / 输出 $0.04。以下为两者的基准测试、延迟与成本全面对比。
切换并对比任意模型
Select any AI models from our verified benchmark database to generate live side-by-side matrices.
Ministral 8B
轻量级高性价比模型Mistral AI 推出的 Ministral 8B 是一款兼顾极低 API 调用成本与高吞吐效率的轻量级大模型，具备 128K 上下文窗口、110 Token/秒生成速度与 38/100 综合智能指数（每百万 Token $0.1 输入 / $0.1 输出）。
$0.10 / 1M
$0.10 / 1M
128K tokens
32.5%
Ministral 3B
轻量级高性价比模型Mistral AI 推出的 Ministral 3B 是一款兼顾极低 API 调用成本与高吞吐效率的轻量级大模型，具备 128K 上下文窗口、180 Token/秒生成速度与 33/100 综合智能指数（每百万 Token $0.04 输入 / $0.04 输出）。
$0.04 / 1M
$0.04 / 1M
128K tokens
22.1%
核心差异一览
- Ministral 3B is 2.5x cheaper for API usage (input + output tokens combined).
- Ministral 3B is faster at 180 tok/s vs 110 tok/s.
- Ministral 8B leads on SWE-bench coding (32.5% vs 22.1%).
| Feature | Ministral 8B | Ministral 3B |
|---|---|---|
Intelligence Index Composite reasoning & knowledge score (0-100) | 38 | 33 |
Provider | Mistral AI | Mistral AI |
Input Price Cost per million input tokens | $0.10 / 1M | $0.04 / 1M |
Output Price Cost per million output tokens | $0.10 / 1M | $0.04 / 1M |
Blended Price Weighted 7:2:1 cache:input:output per 1M | $0.07 / 1M | $0.03 / 1M |
Context Window Maximum input + output tokens | 128K tokens~192 A4 pages of text | 128K tokens~192 A4 pages of text |
Max Output Tokens Per single request | 8K | 8K |
Throughput Speed Tokens generated per second | 110 tok/s | 180 tok/s |
SWE-bench Verified Real-world software engineering | 32.5% | 22.1% |
GPQA Diamond Graduate-level reasoning | 41.5% | 32.5% |
Cached Input Price Prompt caching discount per 1M tokens | $0.05 / 1M | $0.02 / 1M |
各项指标谁更胜一筹？
At-a-glance decision framework权威基准测试得分对比
Official evaluation metrics verified by independent third-party research labs.
Industry benchmark evaluating real-world GitHub issue resolution and multi-file debugging.
Challenging questions created by domain experts to test complex PhD-level reasoning.
Hardened reasoning and academic knowledge benchmark with reduced lucky-guess probability.
Crowdsourced human preference evaluation based on blind, head-to-head model battles.
Extremely difficult, closed-ended questions spanning the frontier of human expert knowledge.
Autonomous multi-step tasks executed inside a real command-line terminal environment.
Research-grade coding problems drawn from real scientific and numerical workloads.
Contamination-free competitive programming problems collected continuously over time.
American Invitational Mathematics Examination — multi-step Olympiad-level reasoning.
Accuracy when reasoning over very large inputs that fill the model context window.
Tokens emitted per second during active generation streaming.
延迟与响应速度
Time to first token (TTFT) and estimated end-to-end response time for a typical 500-token response.
实时 Token 成本模拟器
Adjust sliders to calculate projected monthly cloud API spending with prompt caching discounts.
Ministral 8B · Strengths & Key Advantages
- 提供 128K 上下文窗口与最高 8K 输出容量，核心基准表现优异（SWE-bench Verified 32.5% · GPQA Diamond 41.5%）
- 高达 110 Token/秒的生成吞吐量，首字延迟 (TTFT) 仅约 330 毫秒
- 极具竞争力的 API 定价（每百万 Token $0.1 输入 / $0.1 输出），原生支持提示词缓存、函数调用与结构化 JSON 输出
Considerations & Trade-offs
- 闭源专有云端模型，不支持下载权重进行离线本地物理隔离部署
- 在超长链路的前沿科研推导任务上，深度略逊于顶配重型旗舰推理模型
Ministral 3B · Strengths & Key Advantages
- 提供 128K 上下文窗口与最高 8K 输出容量，核心基准表现优异（SWE-bench Verified 22.1% · GPQA Diamond 32.5%）
- 高达 180 Token/秒的生成吞吐量，首字延迟 (TTFT) 仅约 210 毫秒
- 极具竞争力的 API 定价（每百万 Token $0.04 输入 / $0.04 输出），原生支持提示词缓存、函数调用与结构化 JSON 输出
Considerations & Trade-offs
- 闭源专有云端模型，不支持下载权重进行离线本地物理隔离部署
- 在超长链路的前沿科研推导任务上，深度略逊于顶配重型旗舰推理模型
Token Use & Task Economics
Estimated cost and token consumption to complete one representative agentic task.
| Metric | Ministral 8B | Ministral 3B |
|---|---|---|
Cost per Task Estimated USD to finish one task | $0.08 | $0.03 |
Output Tokens / Task Tokens generated per task | 21K | 15K |
Reasoning Tokens / Task Thinking tokens burned per task | 0K | 0K |
技术规格参数
Architecture, release timeline, capabilities, and modalities.
| Feature | Ministral 8B | Ministral 3B |
|---|---|---|
| Official API Model ID (SDK) | ministral-8b-latest | ministral-3b-latest |
| API Protocol | Mistral AI REST API (OpenAI-Compatible) | Mistral AI REST API (OpenAI-Compatible) |
| Release Date | 2024-10 | 2024-10 |
| Knowledge Cutoff | October 2024 | October 2024 |
| Architecture | Compact Dense Transformer with Interleaved Sliding Window | Ultra-Compact Dense Transformer |
| Parameters | 8B | 3B |
| Enterprise Rate Limits (RPM / TPM) | 5,000 – 15,000 RPM (Production Tier) · 800,000 – 2,000,000 TPM | 5,000 – 15,000 RPM (Production Tier) · 800,000 – 2,000,000 TPM |
| Long-Context Tier (>128K/200K) | Flat rate ($0.1/1M in) across full 128K window | Flat rate ($0.04/1M in) across full 128K window |
| Prompt Cache Policy & TTL | Cached input at $0.05/1M tokens | Cached input at $0.02/1M tokens |
| Computer Use / GUI Agent | API / CLI Tool-Use & DOM Agent ready (no raw pixel desktop click loop) | API / CLI Tool-Use & DOM Agent ready (no raw pixel desktop click loop) |
| MCP & Parallel Tool Calling | Native MCP Client/Server ready · Parallel function calling · 100% Strict JSON Schema | Native MCP Client/Server ready · Parallel function calling · 100% Strict JSON Schema |
| Web Search Grounding & Sandbox | Supported via external Tavily/Exa Search & E2B/Docker Sandbox tools | Supported via external Tavily/Exa Search & E2B/Docker Sandbox tools |
| Official Fine-Tuning Support | API Prompt Engineering & RAG Distillation | API Prompt Engineering & RAG Distillation |
| Multimodal Token Billing | Text BPE (~4 chars/token) | Text BPE (~4 chars/token) |
| Quantization & Min GPU VRAM | Cloud-Hosted Proprietary (FP8 / BF16 Tensor Core Cluster) · Zero Local VRAM Required (100% Serverless Cloud API) | Cloud-Hosted Proprietary (FP8 / BF16 Tensor Core Cluster) · Zero Local VRAM Required (100% Serverless Cloud API) |
| Safety & Frontier Classification | Mistral AI Production Safety & Alignment Verified | Mistral AI Production Safety & Alignment Verified |
| Enterprise Compliance (ZDR/HIPAA/SOC2) | SOC2 Type II · GDPR · Zero Data Retention (ZDR) & HIPAA BAA Available on Enterprise Tier | SOC2 Type II · GDPR · Zero Data Retention (ZDR) & HIPAA BAA Available on Enterprise Tier |
| Input Modalities | text | text |
| Output Modalities | text | text |
Ministral 8B vs Ministral 3B: 常见问题解答 (FAQ)
Which is more intelligent, Ministral 8B or Ministral 3B?
Ministral 8B is more intelligent. It scores 38 compared with 33 on the composite Intelligence Index.
Which model is faster?
Ministral 3B is faster, generating 180 tokens per second compared with 110 tokens per second.
Which model is cheaper to run?
Ministral 3B is cheaper at $0.03 per 1M tokens blended, compared with $0.07 (7:2:1 cache:input:output ratio).
Which model has lower latency?
Ministral 3B has lower latency, with a time to first token of 0.21s compared with 0.33s.
综合结论
综合来看，Ministral 8B（适合：兼顾极低 API 调用成本与高吞吐效率的轻量级任务、企业级 RAG 检索增强生成、自动化智能体工作流与高可靠 API 集成。）与 Ministral 3B（适合：兼顾极低 API 调用成本与高吞吐效率的轻量级任务、企业级 RAG 检索增强生成、自动化智能体工作流与高可靠 API 集成。）各具优势：Ministral 8B 在综合智能与编程基准测试中表现更强，Ministral 3B 拥有更快的 Token 生成速度，而 Ministral 3B 则在规模化 API 部署中具备更高的成本效益。
常见问题 (FAQ)
在编程和深度推理方面，Ministral 8B 和 Ministral 3B 哪个更强？+
在综合智能指数与编程基准评测中，Ministral 8B 表现占优。建议结合上方的 SWE-bench Verified 与 GPQA Diamond 跑分根据具体场景选择。
Ministral 8B 与 Ministral 3B 的 API 价格哪个更便宜？+
Ministral 8B 的 API 价格为输入 $0.10 / 输出 $0.10（每百万 Token），Ministral 3B 为输入 $0.04 / 输出 $0.04。综合混合成本来看，Ministral 3B 更具性价比。
哪款模型拥有更大的上下文窗口和更快的响应速度？+
Ministral 8B 支持 128K Token 上下文窗口，而 Ministral 3B 支持 128K Token。在流式生成速度方面，Ministral 3B 的吞吐表现更快。
Understand key metrics like context window, token pricing, reasoning mode, and SWE-bench scores in our comprehensive AI glossary.
Related reading
- Gemini 1.5 Flash-8B vs Ministral 8B：基准测试、API 价格与速度全面对比 (2026)对比
- Ministral 3B vs Gemini 1.5 Flash-8B：基准测试、API 价格与速度全面对比 (2026)对比
- Aion-3.0 vs Aion-Mini：基准测试、API 价格与速度全面对比 (2026)对比
- Aion-3.0 vs Gemini 3.6 Flash：基准测试、API 价格与速度全面对比 (2026)对比
- Aion-Mini vs Laguna XS 2.1：基准测试、API 价格与速度全面对比 (2026)对比
- Aya Expanse 32B vs Llama 3.2 90B Vision：基准测试、API 价格与速度全面对比 (2026)对比
Benchmark scores, token prices, and rate limits evolve rapidly as AI labs release model updates. Always verify current specifications and pricing on the official provider documentation.
数据来源: Mistral AI 与 Mistral AI 官方 API 文档、模型发布说明及独立基准测试报告 (2026)。