Ministral-3-8B-Instruct-2512
mistralai/Ministral-3-8B-Instruct-2512
Official · FP8 · safetensors
Runtime certified: Transformers vLLM
| Precision / quantization | Format | Bits | Publisher trust | Deployment state |
|---|---|---|---|---|
| FP8 | safetensors | 8 | Official | Runtime certified |
Mistral · Ministral · Ministral 3
Ministral 3 8B Instruct 2512: Specs, GPU requirements, hosting & deployment
Ministral 3 8B Instruct 2512 is a current Ministral release in the Mistral family, certified from its immutable official source revision.
Total parameters8.9B
Active parametersDense model
Context window262,144 tokens
Release statusCurrent · Current
ArchitectureMistral3ForConditionalGeneration
ModalitiesText, Image
LicenseApache License 2.0
Deployment stateDeployable with certified configuration
Commercial stateCommercial Use Allowed
Release date2025-10-31
Architecture, strengths and use cases
Ministral 3 8B Instruct 2512 is an official self-hostable Mistral AI release in the Mistral Ministral track. It uses Mistral3ForConditionalGeneration, carries 8.92B, supports a verified context window of 262,144 tokens, and exposes only source-certified checkpoints and runtime combinations. Its release, license and deployment facts are tied to the official repository revision rather than inferred from the family name.
Ministral 3 8B Instruct 2512 uses Mistral3ForConditionalGeneration with 8.92B. The stored architecture profile records 34 layers, hidden width 4096, 32 attention heads and a 262,144 token context ceiling where the official configuration publishes those fields.
Verified Mistral3ForConditionalGeneration architecture
262,144 token certified context
Official FP8 checkpoint coverage
Private LLM inference
Agentic workflows
Long-context document processing
Vision-language analysis
Certified checkpoints and quantizations
Only exact, source-certified variants attached to this release are shown.
Recommended hardware
Use only the listed transformers, vllm runtime configuration with the exact checkpoint precision. The recommendation engine checks weight memory, KV cache, runtime overhead, safety headroom, GPU architecture and topology before ranking rentable and reference hardware.
Estimated weight memory8.56 GB
KV cache0.53 GB
Runtime overhead6.36 GB
Total with safety margin19.15 GB
Recommended best value
1× NVIDIA H100 SXM 80GB
80 GB total VRAM · nvswitch
60.85 GB headroom · Available
USD 1,860 / monthLowest-cost viable
1× NVIDIA H100 PCIe 80GB
80 GB total VRAM · pcie
60.85 GB headroom · Limited capacity
USD 1,579 / monthHighest performance
1× NVIDIA H200 SXM 141GB
141 GB total VRAM · nvswitch
121.85 GB headroom · Limited capacity
USD 2,345 / monthLicense and self-hosting considerations
The full 8.92B resident weight set must fit across the selected GPUs. Context and concurrency add KV-cache memory; MoE active parameters do not replace resident-weight memory.
Ministral 3 8B Instruct 2512 is published under Apache License 2.0. The release-scoped official license link and verification date are retained with the catalogue record.
Commercial Use Allowed. The official license permits commercial use; its obligations still apply.
Alternatives and sibling variants
Ministral 3 14B Instruct 2512, Ministral 3 3B Instruct 2512, Ministral-3-14B-Instruct-2512
Release timeline
mistral-ai/mistral/ministral-3-8b-instruct-2512 @ 2db7feae217eadd1ceeb051f5db5287b975d5786155df54bba7b44137234c275
Frequently asked questions
How much VRAM does Ministral 3 8B Instruct 2512 require?
The exact estimate depends on checkpoint precision, context and concurrency; the page recommendation uses the certified memory engine and compatible runtime configurations.
Which runtimes support Ministral 3 8B Instruct 2512?
Certified upstream runtime evidence is available for transformers, vllm.
Can Ministral 3 8B Instruct 2512 be used commercially?
The stored commercial classification is Commercial Use Allowed. Consult the official Apache License 2.0 terms for the final obligations.
Official references
Last verified: 2026-08-26