AI Model Comparison
Side-by-side comparison of pricing, context window, benchmarks, capabilities, and license. Every figure comes from the developer's own pages, last checked on ; each model's page lists its sources. A figure we could not source shows as a dash.
Model A — Preferred Networks
PLaMo 3.0 Prime is the flagship of Preferred Networks’ PLaMo family, developed from scratch in Japan and released on June 22, 2026. It is offered as a reasoning model and a faster non-reasoning model, through an API or on-premises, with its context extended from 64K to 256K tokens.
Model B — Sarvam AI
Sarvam-105B is a mixture-of-experts reasoning model from Sarvam AI, released on March 6, 2026, trained from scratch on 12 trillion tokens with multi-head latent attention, for reasoning, coding and agentic tasks, with a focus on Indian languages.
| Property | PLaMo 3.0 Prime | Sarvam-105B |
|---|---|---|
| Released | Jun 22, 2026 | Mar 6, 2026 |
| Developer | Preferred Networks | Sarvam AI |
| Type | reasoning | reasoning |
| License | proprietary | open-weight |
| Context window | — | — |
| Parameters | — | 105B+ total (MoE, 128 experts); 10.3B active |
| Developer status | — | — |
| Input $/M tokens | — | — |
| Output $/M tokens | — | — |
Bold green values indicate the winner on that dimension where comparison is meaningful (lower price wins; bigger context wins; more open license wins; newer date wins). Prices are the ones each developer lists today; a model it no longer prices has none.
No exclusive capabilities
Programmatic access available at /api/ai/compare?a=plamo-3-prime&b=sarvam-105b. CORS is open. Citation guidance: /about/editorial-process.