K-EXAONE-2.0-750B-A37B
LGAI-EXAONE/K-EXAONE-2.0-750B-A37B
Official · BF16 · safetensors
Runtime certified: Transformers vLLM SGLang
| Precision / quantization | Format | Bits | Publisher trust | Deployment state |
|---|---|---|---|---|
| BF16 | safetensors | 16 | Official | Runtime certified |
EXAONE · General · K-EXAONE 2.0
K-EXAONE 2.0 750B-A37B: Specs, GPU requirements, hosting & deployment
K-EXAONE 2.0 750B-A37B is a current General release in the EXAONE family, certified from its immutable official source revision.
Total parameters750.0B
Active parameters37.0B
Context window262,144 tokens
Release statusCurrent · Current
ArchitectureExaoneMoeForCausalLM
ModalitiesText
Licenseapache-2.0
Deployment stateDeployable with certified configuration
Commercial stateCommercial Use Allowed
Release date2026-07-29
Architecture, strengths and use cases
K-EXAONE 2.0 750B-A37B is an official self-hostable LG AI Research release in the EXAONE General track. It uses ExaoneMoeForCausalLM, carries 750B / 37B MoE, 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.
K-EXAONE 2.0 750B-A37B uses ExaoneMoeForCausalLM with 750B / 37B MoE. The stored architecture profile records 78 layers, hidden width 6144, 64 attention heads and a 262,144 token context ceiling where the official configuration publishes those fields.
Verified ExaoneMoeForCausalLM architecture
262,144 token certified context
Official BF16, FP8 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.
K-EXAONE-2.0-750B-A37B-FP8
LGAI-EXAONE/K-EXAONE-2.0-750B-A37B-FP8
Official · FP8 · safetensors
Runtime certified: vLLM SGLang
| Precision / quantization | Format | Bits | Publisher trust | Deployment state |
|---|---|---|---|---|
| FP8 | safetensors | 8 | Official | Runtime certified |
Recommended hardware
Use only the listed transformers, vllm, sglang 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 memory1438.89 GB
KV cache2.44 GB
Runtime overhead316.08 GB
Total with safety margin2179.19 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 750B 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.
K-EXAONE 2.0 750B-A37B is published under apache-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
K-EXAONE 236B-A23B, K-EXAONE-2.0-750B-A37B-FP8, K-EXAONE-2.0-750B-A37B-NVFP4
Release timeline
lg-ai-research/exaone/k-exaone-2-750b-a37b
Source details
Source revision: 863ab927f6da1d3740ee17b121d4cbd5f2acecd44c1366051fe6e5ff24f63260
Frequently asked questions
How much VRAM does K-EXAONE 2.0 750B-A37B 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 K-EXAONE 2.0 750B-A37B?
Certified upstream runtime evidence is available for transformers, vllm, sglang.
Can K-EXAONE 2.0 750B-A37B be used commercially?
The stored commercial classification is Commercial Use Allowed. Consult the official apache-2.0 terms for the final obligations.
Official references
Last verified: 2026-09-03