From the NVIDIA edition of October 13, 2026
LINAGORA: Luciole open models adapt the Nemotron recipe, trained on Jean Zay
LINAGORA / OpenLLM France · Software vendor and open-model consortium · France
- Publisher
- Carnet souverain (Michel-Marie Maudet blog, reposting LINAGORA)
- Original publication
- 2026-07-08
- Source retrieved
- 2026-10-14
- Event date
- 2026-07-08
What happened
The post says LINAGORA and OpenLLM France build open multilingual models with GENCI and AI Factory France support on the Jean Zay supercomputer. Luciole comes in 1B, 8B (hybrid Mamba-Transformer) and 23B versions, all Apache 2.0, with checkpoints, training data and code published and a corpus of around 5 trillion tokens. All Luciole architectures are adapted from Nemotron, and model cards credit a custom adaptation of the Nemotron 3 Nano recipe and training with NeMo on Jean Zay H100s. Twake.ai is listed in the same open-source ecosystem.
Why it matters
Moderate for state AI offices and universities considering open models built on public compute; explains the Nemotron link behind NVIDIA's Twake.ai claim.
Evidence and measured results
Vendor-affiliated description; no benchmarks.
Limitations and uncertainty
Why now: it is the documented basis for NVIDIA's Oct 13 'built on Nemotron' attribution; it does not say which models Twake.ai deployments run.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-14; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Public-sector buyers need to know what 'built on' a vendor's model actually means for provenance and control.
- Stakeholders
- State AI governance offices
- University AI research groups
- Public-sector procurement officers
- Discovery questions
- Do you require model provenance documentation?
- Would open training data and checkpoints satisfy your audit needs?
- Do you have public compute for fine-tuning?
- Value hypothesis
- Interpretation: fully open models with documented lineage can ease audits; the post reports no task benchmarks.
- Potential engagement
- Bounded model-provenance review for a public AI deployment, distinguishing training recipe, datasets and runtime model.
- Unsupported claims
- That Twake.ai deployments run Luciole or Nemotron
- Benchmark performance
- Production readiness of foundation models
Pre-sales engineering
Role takeaway
- Fit
- Moderate for research universities and state AI offices interested in open models trained on public compute.
- Architecture and integration
- Luciole 1B/8B/23B; hybrid Mamba-Transformer 8B; trained with NeMo on Jean Zay H100s.
- Prerequisites
- Fine-tuning compute
- Evaluation datasets
- Constraints
- Foundation models, not chat assistants
- European-language focus
- Security
- Open data and code aid auditing; no security evaluation reported.
- Proof of value
- Evaluate a Luciole model on a representative public-sector task
- Trace dataset licenses
Delivery
Role takeaway
- Work
- Fine-tune and evaluate open models for a specific public use case.
- Dependencies
- Public compute allocation
- Evaluation team
- Ownership
- Research or AI platform team.
- Skills and adoption
- Requires ML engineering skills for fine-tuning and evaluation.
- Governance checkpoints
- Model provenance documented
- Evaluation completed before deployment
- Acceptance criteria
- Proposed: task accuracy measured against a baseline model
- Risks
- Overstating vendor involvement
- Insufficient evaluation
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Nemotron-derived architectures and NeMo tooling; deployable where data stays under the operator's control.
Governance
Who approves, reviews and stays accountable for outcomes?
Openness beyond weights supports auditability.
Security and privacy
What data, permissions and controls need testing?
Not addressed.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Public compute access lets regional builders participate.
Procurement
What should contracts, pricing and exit terms secure?
Apache 2.0 models reduce license constraints.
Operating model
Which teams own the service once it runs?
Regional model builders depend on public supercomputing allocations.
Publication history
- 2026-10-13NVIDIA · Issue 387 resources
Stable resource ID: linagora-luciole-nemotron-openllm-france-20260708