Compare plans

Choose the plan that fits your needs

Free Trial

Try every Solo Dual AI feature for 48 hours — no payment required.

₩0 48 hours
  • Full Solo Dual AI experience
  • Unlimited local AI
  • Gemini & Claude cloud AI (trial)
  • No credit card needed

Non-AI

No AI

File management, viewer, RAG search, and citation analysis — without AI.

₩39,900 /mo
  • Unlimited documents
  • Unlimited documents
  • Keyword + semantic hybrid search
  • Citation analysis · source tracing
Recommended

Solo Local AI

Local AI Only

Legal AI that runs entirely on your PC — no internet needed.

₩69,900 /mo
  • Unlimited local AI
  • No Cloud AI
Recommended

Solo Cloud AI

Cloud AI Only

Powerful cloud-based legal AI — no high-end machine required.

₩79,900 /mo
  • No local AI
  • Gemini & Claude cloud AI

Solo Dual AI

Local AI Cloud AI

Local and cloud AI combined — always ready whether online or offline.

₩109,900 /mo
  • Unlimited local AI
  • Gemini & Claude cloud AI

Local AI Performance & Requirements

Running local AI requires at least 6 GB of disk space and 16 GB of RAM.

Gemma 4 E4B

Lightweight, efficient — runs smoothly even on lower-spec Macs

  • File size 5.4 GB (Q4_K_M)
  • Disk space 6 GB+
  • Parameters 8B
  • Context 128K tokens
  • Min RAM 8 GB
  • Rec. RAM 16 GB+

Gemma 4 26B-A4B

MoE architecture — 26B parameter quality at near-7B inference speed

  • File size 21.3 GB (Q5_K_XL)
  • Disk space 22 GB+
  • Parameters 26B trained / 4B active (MoE)
  • Context 256K tokens
  • Min RAM 24 GB
  • Rec. RAM 32 GB+

Benchmarked on Apple Silicon (M-series). Unified Memory is used as VRAM — no discrete GPU required.

Performance Benchmark Comparison

Benchmark Gemma 4 E4BGemma 4 26B-A4BGemini 2.5 FlashGemini 3.1 Flash
MMLU Pro Expert knowledge 69.4%~83%88.4%
GPQA Diamond Science & reasoning 79.2%82.8%86.9%
AIME 2026 Math competition 42.5%88.3%88.0%
LiveCodeBench v6 Coding ability 77.1%~65%72.0%
RULER 128K Long-doc retrieval 66.4%
Context window Max input length 128K256K1M1M

Sources: Google DeepMind Gemma 4 official release, Artificial Analysis, BenchLM (Apr 2026). Bold = best score in row. '—' = not reported.

Cloud AI Billing

Your subscription includes monthly units. After they run out, choose whichever billing method is more cost-effective for you.

Solo Cloud AI

₩79,900/mo

600 units

Monthly cloud AI included

  • Gemini 2.5 Flash 600 calls
  • Gemini 3.0 Flash 150 calls
  • Gemini 2.5 Pro 75 calls
  • Gemini 3.1 Pro ~46 calls
  • Claude Haiku 4.5 ~85 calls
  • Claude Sonnet 4.6 30 calls
  • Claude Opus 4.6 ~18 calls

1 unit = 1 Gemini 2.5 Flash call. Higher-capability models consume more units.

Solo Dual AI

₩109,900/mo

800 units + unlimited local AI

Monthly cloud AI included

  • Gemini 2.5 Flash 800 calls
  • Gemini 3.0 Flash 200 calls
  • Gemini 2.5 Pro 100 calls
  • Gemini 3.1 Pro ~61 calls
  • Claude Haiku 4.5 ~114 calls
  • Claude Sonnet 4.6 40 calls
  • Claude Opus 4.6 ~24 calls

1 unit = 1 Gemini 2.5 Flash call. Higher-capability models consume more units.

After units run out — choose whichever is more favorable

Pick per-call or per-token billing — whichever suits you better. Frequent short queries favor per-token; if you prefer predictable costs, per-call is the better choice.

Per-call billing

Billed per request — easy to predict costs

  • Gemini 2.5 Flash ₩50/call
  • Gemini 3.0 Flash ₩70/call
  • Gemini 2.5 Pro ₩210/call
  • Gemini 3.1 Pro ₩280/call
  • Claude Haiku 4.5 ₩130/call
  • Claude Sonnet 4.6 ₩380/call
  • Claude Opus 4.6 ₩650/call
OR

Per-token billing

Charged only for tokens used — better for short queries

Model Input Output
Gemini 2.5 Flash ₩1/1K ₩5/1K
Gemini 3.0 Flash ₩3/1K ₩13/1K
Gemini 2.5 Pro ₩6/1K ₩42/1K
Gemini 3.1 Pro ₩9/1K ₩50/1K
Claude Haiku 4.5 ₩5/1K ₩21/1K
Claude Sonnet 4.6 ₩13/1K ₩63/1K
Claude Opus 4.6 ₩21/1K ₩104/1K

e.g. Gemini Flash, 1 query (2K input + 0.5K output tokens) → ~₩5

Prices include 10% VAT; save 10% with annual billing.