Kimi-K3
Official · MXFP4 · safetensors
Runtime certified: Transformers
| Precision / quantization | Format | Bits | Publisher trust | Deployment state |
|---|---|---|---|---|
| MXFP4 | safetensors | 4 | Official | Runtime certified |
Kimi · General · Kimi K3
Kimi K3: Specs, GPU requirements, hosting & deployment
Kimi K3 is a current General release in the Kimi family, certified from its immutable official source revision.
Total parameters2,800.0B
Active parameters100.0B
Context window1,048,576 tokens
Release statusCurrent · Current
ArchitectureKimiK3ForConditionalGeneration
ModalitiesText
LicenseKimi K3 License
Deployment stateDeployable with certified configuration
Commercial stateCommercial Use Allowed With Conditions
Release date2026-06-13
Architecture, strengths and use cases
Kimi K3 is an official self-hostable Moonshot AI release in the Kimi General track. It uses KimiK3ForConditionalGeneration, carries 2800B / 100B MoE, supports a verified context window of 1,048,576 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.
Kimi K3 uses KimiK3ForConditionalGeneration with 2800B / 100B MoE. The stored architecture profile records 93 layers, hidden width 7168, 96 attention heads and a 1,048,576 token context ceiling where the official configuration publishes those fields.
Verified KimiK3ForConditionalGeneration architecture
1,048,576 token certified context
Official MXFP4 checkpoint coverage
Private LLM inference
Agentic workflows
Long-context document processing
Certified checkpoints and quantizations
Only exact, source-certified variants attached to this release are shown.
Recommended hardware
Use only the listed transformers 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 memory1342.97 GB
KV cache5.04 GB
Runtime overhead299.28 GB
Total with safety margin2042.64 GB
Technically compatible reference sizing
No current FusionPointAI inventory passes every runtime, architecture, topology and memory gate for this exact checkpoint. Use the estimator for global compatible hardware.
License and self-hosting considerations
The full 2800B 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.
Kimi K3 is published under Kimi K3 License. The release-scoped official license link and verification date are retained with the catalogue record.
Commercial Use Allowed With Conditions. Commercial use is conditional. Review the linked official terms before deployment.
Alternatives and sibling variants
Kimi K2.6, Kimi K2.5, Kimi-K3
Curated comparisons
Release timeline
moonshot/kimi/kimi-k3 @ a2873a31d078c3b0a6dca6eca78e737dda9f3349d85ad04bc5aef21e2e002742
Frequently asked questions
How much VRAM does Kimi K3 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 Kimi K3?
Certified upstream runtime evidence is available for transformers.
Can Kimi K3 be used commercially?
The stored commercial classification is Commercial Use Allowed With Conditions. Consult the official Kimi K3 License terms for the final obligations.
Official references
Last verified: 2026-08-26