Mistral Small 3.1
Mistral Small 3.1 is a chat-oriented language model from Mistral in the company's Small line, positioned as a compact option below its larger Mistral models.
API Pricing
| Provider | Input / 1M | Output / 1M |
|---|---|---|
| $0.351 | $0.555 |
Prices updated daily. Last check: Oct 16, 2026
Compare API pricing for every Mistral model →Mistral Small 3.1 pricing by provider
Input, output, and batch rates, plus alternatives, for one provider at a time.
Performance & Benchmarks
Source: Artificial Analysis →Reasoning & Knowledge
- MMLU-Pro65.9%
- GPQA Diamond45.4%
- Humanity's Last Exam4.3%
Coding
- LiveCodeBench21.2%
- SciCode27.8%
Math
- AIME 20253.7%
- AIME9.3%
- MATH-50070.7%
Agentic & Tool Use
- Terminal-Bench Hard7.6%
- Terminal-Bench v2.126.2%
- τ²-bench25.1%
- τ-bench Banking7.4%
Instruction & Long Context
- IFBench29.9%
- Long-Context Reasoning22.3%
Benchmarks measured Oct 2026. Scores are independent evaluations, not vendor-reported.
Model Details
General
- Creator
- Mistral
- Modalities
- Text
Capabilities
- Open Source
- Yes
Strengths & Limitations
Strengths
- Positioned in Mistral's Small tier, which targets lower-cost, higher-volume chat workloads rather than maximum capability
- Iteration on the earlier Mistral Small 3 release within the same family, so prompts and integrations built for the line generally carry over
- Compact models in this class are typically well suited to latency-sensitive assistant and extraction tasks
- Mistral maintains a broad family (Small, Medium, Large, Magistral), making tier-to-tier upgrades straightforward if a workload outgrows Small
- Offered through multiple inference providers, so teams can compare hosts on the pricing table rather than being tied to a single endpoint
- Standard chat completion interface makes it usable as a drop-in default for routine text traffic
Limitations
- We do not have a confirmed context window recorded for this model — verify with your provider before relying on long inputs
- No verified benchmark scores are tracked in our database for Mistral Small 3.1, so capability comparisons must come from third-party evaluations
- Throughput and time-to-first-token measurements are not populated in our data, making speed comparisons unavailable here
- As a Small-tier model, it is not intended to match Mistral's larger models on complex reasoning or long-horizon agentic tasks
- Feature support such as tool calling or image input may vary by host and is not confirmed in our metadata
Key Features
About Mistral Small 3.1
Common Use Cases
Mistral Small 3.1 fits workloads where a compact chat model is sufficient: customer-facing assistants handling routine questions, summarization of documents and transcripts, structured data extraction from unstructured text, classification and routing, drafting and rewriting content, and serving as the first-pass model in a tiered setup that escalates harder requests to a larger Mistral model. Because Small-tier models are generally chosen for cost-per-request and response speed, they suit high-request-volume production traffic more than deep multi-step reasoning, long research tasks, or complex agentic pipelines, where a larger tier is typically the better match. Teams building on it should validate context length and any tool-calling requirements against their chosen provider.
Frequently Asked Questions
How much does Mistral Small 3.1 cost?
Pricing depends on which provider serves the model and on the pricing type — input and output tokens are usually billed at different rates, and some hosts offer batch or cached-input discounts. Check the pricing table on this page for current per-provider rates.
What is Mistral Small 3.1 best used for?
It is best suited to high-volume chat and text-processing work: assistant responses, summarization, extraction, classification, and drafting. As a Small-tier model it is typically chosen when cost and latency matter more than maximum reasoning depth.
How does it differ from Mistral Small 3?
Mistral Small 3.1 is a later iteration in the same Small line. Both target the same compact tier, so integrations built for Small 3 generally transfer, but you should re-run your own evaluations rather than assuming identical behavior.
What context window does Mistral Small 3.1 support?
We do not have a confirmed context window recorded for this model in our database. Context length can also be capped differently by individual hosts, so confirm the limit with the provider you plan to use.
Should I use Mistral Small 3.1 or a larger Mistral model?
Use the Small tier for routine, high-frequency requests where throughput and per-request cost dominate. Move to a larger Mistral tier when tasks involve multi-step reasoning, long documents, or agentic workflows that a compact model handles less reliably.