Opper AI partners with Nextbit to open a Spanish route on the gateway

Stockholm, Sweden, August 2026. Opper AI and Nextbit are partnering to bring Nextbit's Spanish inference to the Opper AI gateway. Builders can now reach Nextbit's catalogue through the Opper API, from Gemma 4 26B A4B and Qwen3 14B to ALIA 40B, the model the Barcelona Supercomputing Center built for Spain's public administration, with more on the way.

Why this matters

The first Spanish route on the gateway. Opper already routes to hosts in Sweden, Italy, Germany, France, Finland, Norway, and Ireland. Nextbit adds Spain, and it does so from infrastructure the company owns rather than rents: Nextbit built its first HPC data center in 2022 and its first AI-dedicated facility in Spain in 2024. For teams whose procurement asks where a call is physically processed, that is a shorter answer than a chain of subcontractors.

A national model on national infrastructure. ALIA 40B is the instruction-tuned release of Spain's public language model, trained from scratch by the Language Technologies Lab at the Barcelona Supercomputing Center on MareNostrum 5, the EuroHPC pre-exascale machine BSC hosts. It saw 9.83 trillion tokens across 35 European languages, with post-training concentrated on Spanish, Catalan, Basque, Galician, and English, and it ships under Apache 2.0 with the training scripts and configurations published alongside the weights. It is funded by Spain's Ministry for Digital Transformation and supported by the Catalan government's Aina project, which makes it one of the few frontier-scale models built as public infrastructure rather than a product. Serving it from a Spanish data center closes a loop that usually stays open.

Serverless or dedicated, on the same platform. Nextbit offers pay-per-token serverless endpoints for experimentation and variable traffic, and single-tenant dedicated GPU deployments for teams that need guaranteed throughput or want to run their own fine-tuned or private weights. The route Opper consumes is the Spanish one, and the full residency, retention, and GDPR detail per route lives on the Nextbit provider page.

Clear data handling. Nextbit does not train, fine-tune, or otherwise adapt any model on customer data. Data attached to API requests may be held for up to 90 days in isolated, access-restricted environments for diagnosing and improving service reliability, then deleted automatically. An Article 28 GDPR data processing agreement is available on request, and zero data retention can be arranged through Opper Enterprise.

More than a router: the AI control plane. Routing to Nextbit is just the entry point. Every call through Opper runs on its AI control plane: intelligent routing across models and regions, full observability, real-time PII masking and content filtering, budget caps, and audit trails. Pin Nextbit for a task or set it as a fallback, and because Nextbit is OpenAI-compatible, getting there is a model string, not a migration.

"Spain has been building serious AI infrastructure for years, and ALIA is proof of what comes out of it. Nextbit gives our developers a route into that, open models served from a Spanish data center, with the national model among them."

— Göran Sandahl, Co-founder and CEO, Opper AI

"We built Nextbit so teams in Spain and across Europe can run open models on infrastructure we operate ourselves, without stitching together a provider for compute, another for models, and another for everything in between. Opper puts that in front of a wide community of builders through a single API."

— Marc Sanmillan, CEO, Nextbit

Models live today

The catalog below is fetched live from Opper's model API and filtered to Nextbit-hosted models. Availability, context windows, and pricing stay in sync with what's actually callable through Opper. All Nextbit routes on Opper are hosted in Spain.

Alongside the open-weight staples, the catalog carries a set of community fine-tunes from independent authors, Euryale v2.3, MythoMax L2, ReMM SLERP L2, and UnslopNemo. These are the character and creative-writing models that usually only run on US hosts, and on Nextbit they run inside the EU.

Get started

Paste this into your coding agent (Claude Code, Cursor, Codex, and more) and it will set up Opper and route to Nextbit for you:

Use curl to download, read and follow: https://skills.opper.ai
Then set up Opper to use Nextbit as the provider, e.g. nextbit/gemma4:26b-a4b.

Prefer a direct call? Opper is drop-in compatible with the OpenAI, Anthropic, and Google SDKs, so one API key and the model string are all you need:

OpenAIAnthropicGeminicURL

1import OpenAI from "openai";
2
3const client = new OpenAI({
4  apiKey: process.env.OPPER_API_KEY,
5  baseURL: "https://api.opper.ai/v3/compat",
6});
7
8const completion = await client.chat.completions.create({
9  model: "nextbit/gemma4:26b-a4b",
10  messages: [{ role: "user", content: "Hello" }],
11});
12console.log(completion.choices[0].message.content);

Follow the quick start in our docs for evaluations, fallbacks, and structured output.

About Nextbit

Nextbit (NEXTBIT 256, S.L.) is a Spanish AI infrastructure company founded in 2021 and backed by the Valencia accelerator Lanzadera. It serves open-weight models through an OpenAI-compatible API from its own data center in Spain, alongside dedicated single-tenant GPU endpoints for custom, fine-tuned, and private models. Nextbit does not use customer data to train or fine-tune any model.

About Opper AI

Opper AI is the European AI gateway and control plane for agents: one EU-hosted, GDPR-compliant API across 700+ models, with smart routing, automatic fallbacks, built-in evaluations and observability, real-time guardrails, and full OpenAI SDK compatibility.