Side-by-side comparison of LFM2.5-8B-A1B (Liquid AI) and Mistral 7B Instruct v0.2 (Mistral AI · France) for self-hosted deployment of the open-weight model. LFM2.5-8B-A1B is rated conditional; Mistral 7B Instruct v0.2 is conditional. They part ways on licence: LFM2.5-8B-A1B is "LFM Open License 1.0", Mistral 7B Instruct v0.2 is "Apache 2.0".
| Field | ||
|---|---|---|
| Summary | ||
| Verdict | Conditional Per the published LFM Open License v1.0, Liquid AI's terms are Apache-2.0-derived but cap free commercial use at USD 10M annual company revenue — above that threshold deployers must negotiate a paid licence with [email protected]. The model itself is a strong 8.3B / 1.5B-active MoE built for CPU and edge inference, but the 38-trillion-token training corpus is undisclosed and Liquid AI is a US-based MIT spinoff, so EU regulated buyers should run their own DPIA and document the data-transfer story alongside any commercial-tier negotiation. | Conditional Based on published licence terms, Mistral 7B Instruct v0.2 is an EU-origin open-weight model under standard Apache 2.0 — commercial deployment and self-hosting are permitted without field-of-use restrictions. Training-data opacity is the primary EU AI Act Art. 53 gap, but the French controller, absence of CLOUD Act exposure, and Mistral's published DPA make this a strong baseline for regulated EU workloads. |
| Last reviewed | 2026-05-30 | 2026-04-17 |
| Open-weight | ||
| Licence | LFM Open License 1.0 | Apache 2.0 |
| Commercial use | Free under $10M revenue | Unrestricted |
| Training data | Undisclosed | Undisclosed |
| Origin | USA (Boston) | EU (France) |
| Performance & pricing? | ||
| Quality index | — | 7/100 |
| Speed | — | 193 tok/s |
| Blended price | — | $0.25/M |
| Context window | — | — |
| Evidence | ||
| Sources | ||
No overlapping sources between the two entries.