Stockmark vs ELYZA
A side-by-side look at scores, pricing and features — with RECATOOLS' ASEAN-aware verdict for each.
|
|
ELYZA
Japan's Japanese-first LLM — from U Tokyo's Matsuo Lab, backed by KDDI
Visit
|
|
|---|---|---|
| RECATOOLS Score | 6.4 / 10 | 7.4 / 10 |
| Capability | ||
| Value for money | ||
| Ease of use | ||
| ASEAN readiness | ||
| API quality | ||
| Pricing | Enterprise | Enterprise |
| Free tier | — | Free demo ended May 2026; open-weight models freely downloadable on Hugging Face |
| Paid from | — | Enterprise contract only — no public pricing; contact sales via elyza.ai |
| Has API | ✓ | ✓ |
| Open source | ✗ | ✗ |
| Free to use | ✗ | ✗ |
| Users | — | ~110K users in 11 days at ELYZA Pencil launch (2022); enterprise B2B focus since |
| Founded | — | 2018 |
| Maker | — | KDDI Corporation |
| Verdict | Stockmark is a legitimate enterprise AI vendor rather than a chatbot wrapper: its Stockmark-2-100B family is trained from scratch specifically to suppress hallucination on dense Japanese business documents, and adoption... |
ELYZA has the cleanest academic-to-enterprise pipeline among Japan's domestic LLM shops: University of Tokyo Matsuo Lab research, a 2024 KDDI majority buy-in (53.4% combined stake), and production deployments that aren't... |
| Full review → | Full review → |
← Back to AI Directory
Comparisons cover up to 4 tools. Scores are RECATOOLS editorial assessments; verify current pricing on each vendor's site.