LG AI Research · څو-مودالي

د EXAONE ماډلونه: نسخې، ځانګړنې، د GPU اړتیاوې او کوربه‌توب

LG AI Research EXAONE combines multilingual reasoning and vision-language checkpointونه with explicit deployment evidence for large sparse releases.

اوسنۍ نسخې د انجنیري خدمتونه وپلټئ ماډلونه پرتله کړئ

Technical deployment reference illustration for EXAONE — Reference deployment flow; exact architecture facts are listed separately.

لنډه کتنه او مخینه

EXAONE is an official LG AI Research model family represented here through independently verified release, checkpoint, license, د اجرا چاپېریال and hardware evidence. LG AI Research EXAONE combines multilingual reasoning and vision-language checkpointونه with explicit deployment evidence for large sparse releases. Current public releases are grouped by the General, Vision language tracks so sibling sizes and specialist branches are not presented as false chronological generations.

FusionPointAI currently tracks 10 certified public EXAONE releases. The catalogue preserves release dates and evaluates Current, Recent, Previous and Legacy status within each model track rather than by database insertion order.

Multilingual reasoning

Vision-language track

Large-model deployment evidence

معماري او د ماډل لړۍ

The verified public range spans 33B to 750B total parameters and context windows up to 262,144 ټوکنونه. Each release states whether it is dense or mixture-of-experts and records active parameters separately when the official source publishes them.

عامې نسخې10

د ماډل لړۍ2

اوسنی نسلK-EXAONE 2.0 750B-A37B

وروستی تایید2026-09-03T00:00:00.000000Z

نسلونه، لړۍ او بڼې

عمومي

K-EXAONE

وروستی نسل

K-EXAONE 236B-A23B

۲۳۶٫۰B پارامېټرونه · ۲۶۲٬۱۴۴ context

MoE ExaoneMoeForCausalLM ناڅرګند / د سپارښتنې وړ نه دی

k-exaone-236b-a23b

وروستی نسل د 2 هم‌نسله بڼې

K-EXAONE-236B-A23B-GGUF

جوړوونکي دا معلومات نه دي خپاره کړي. پارامېټرونه · N/A / نېغه پرتله نه شي کېدای context

Quantization ناڅرګند / د سپارښتنې وړ نه دی

K-EXAONE-236B-A23B-FP8

جوړوونکي دا معلومات نه دي خپاره کړي. پارامېټرونه · N/A / نېغه پرتله نه شي کېدای context

Quantization ناڅرګند / د سپارښتنې وړ نه دی

k-exaone-2-0

اوسنی نسل د 3 هم‌نسله بڼې

K-EXAONE-2.0-750B-A37B-FP8

جوړوونکي دا معلومات نه دي خپاره کړي. پارامېټرونه · N/A / نېغه پرتله نه شي کېدای context

Quantization

K-EXAONE-2.0-750B-A37B-NVFP4

جوړوونکي دا معلومات نه دي خپاره کړي. پارامېټرونه · N/A / نېغه پرتله نه شي کېدای context

Quantization

K-EXAONE 2.0 750B-A37B

۷۵۰٫۰B پارامېټرونه · ۲۶۲٬۱۴۴ context

مخکښ · MoE ExaoneMoeForCausalLM

لید او ژبه

exaone-4-5

اوسنی نسل د 4 هم‌نسله بڼې

EXAONE-4.5-33B-AWQ

جوړوونکي دا معلومات نه دي خپاره کړي. پارامېټرونه · N/A / نېغه پرتله نه شي کېدای context

Quantization ناڅرګند / د سپارښتنې وړ نه دی

EXAONE-4.5-33B-FP8

جوړوونکي دا معلومات نه دي خپاره کړي. پارامېټرونه · N/A / نېغه پرتله نه شي کېدای context

Quantization ناڅرګند / د سپارښتنې وړ نه دی

EXAONE-4.5-33B-GGUF

جوړوونکي دا معلومات نه دي خپاره کړي. پارامېټرونه · N/A / نېغه پرتله نه شي کېدای context

Quantization ناڅرګند / د سپارښتنې وړ نه دی

EXAONE 4.5 33B

۳۳٫۰B پارامېټرونه · ۲۶۲٬۱۴۴ context

لید · Dense Exaone4_5_ForConditionalGeneration ناڅرګند / د سپارښتنې وړ نه دی

د ځای‌پرځای کولو ایکوسیستم

Choose a specific EXAONE release and certified checkpoint before sizing hardware. FusionPointAI rejects unknown د اجرا چاپېریال configurations, calculates resident memory from total model weights, then distinguishes globally compatible GPUs from currently rentable inventory.

Validate the exact checkpoint license, د اجرا چاپېریال version, precision kernel, context length, concurrency and multi-GPU topology. Active MoE parameters affect compute, but resident model memory continues to use total stored weights.

عامې کارونې

Multilingual assistants

Document reasoning

Vision-language analysis

د جواز لنډه کتنه

Licenses are certified per release, not inferred from the provider. Review the release page before commercial deployment because terms can differ between generations and variants.

Commercial-use status is a source-backed catalogue classification, not legal advice. Follow the linked official license and attribution requirements.

د نسخو مهالوېش

lg-ai-research/exaone/exaone-45-33b-fp8

م. ۲۰۲۶ سېپتمبر ۳

Which EXAONE release is current?

Current releases are K-EXAONE 2.0 750B-A37B, EXAONE 4.5 33B, evaluated within their respective tracks.

Can EXAONE be self-hosted?

Only releases with a verified official checkpoint and supported د اجرا چاپېریال configuration are presented as deployable.

Which GPU should run EXAONE?

GPU selection depends on the exact variant, checkpoint precision, context length and concurrency; use the release hardware recommendations rather than the family name alone.

رسمي مراجع

وروستی تایید: 2026-09-03