Opper AI partners with Sference to add European inference for frontier open models
By Felix Wunderlich -
Stockholm, Sweden, August 2026. Opper AI and Sference are partnering to bring Sference's managed inference for open models to the Opper AI gateway, on a European route with compute inside the EEA. Builders can now reach Kimi K3, GLM-5.2, DeepSeek V4 Flash, Qwen 3.6 35B-A3B, Qwen3-VL 30B, and BottleCap AI's ThinkingCap through the Opper API, with zero data retention on by default.
Why this matters
A European route for frontier open weights. Sference runs inference and storage of customer content on infrastructure inside the EEA, zero data retention is on by default rather than an upsell, there is no training on customer data, and the DPA is published rather than negotiated. The full privacy detail lives on the Sference provider page.
A focused catalogue that goes up to the frontier. The roster is curated rather than long-tail: Moonshot's Kimi K3 with its full one-million-token context window, tool calling, structured output, and adjustable reasoning effort, alongside GLM-5.2, DeepSeek V4 Flash, the efficient Qwen 3.6 line, and BottleCap AI's reasoning-tuned ThinkingCap, all with the same residency and retention guarantees.
Built with production discipline. Every model is pinned to a specific version, so the checkpoint you evaluated is the checkpoint that serves your traffic. Realtime endpoints, discounted flex processing, and async completion windows run behind one OpenAI-compatible API on the same GPU fleet, so interactive agent turns and background jobs share the same models and the same audit trail.
More than a router: the AI control plane. Routing to Sference 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 Sference for a task or set it as a fallback, and because Sference is OpenAI-compatible, getting there is a model string, not a migration.
"We're glad to welcome Sference to Opper. They give our developers exactly what European teams keep asking for, frontier open models served in Europe with zero data retention by default."
— Göran Sandahl, Co-founder and CEO, Opper AI
"We built Sference so teams can run open models in production without operating the infrastructure themselves, on a platform built and run in Europe. Opper puts that in front of a large community of builders through one API, and we look forward to what they ship with it."
— Jernej Štrasner, Co-founder and CEO, Sference
Models live today
The catalog below is fetched live from Opper's model API and filtered to Sference-hosted models. Availability, context windows, and pricing stay in sync with what's actually callable through Opper. All Sference routes are European, with compute inside the EEA.
| Model | Region | Context | Input / 1M | Output / 1M |
|---|---|---|---|---|
| sference/Qwen/Qwen3-VL-30B-A3B-Instruct | EU | 262K | $0.40 | $2.00 |
| sference/Qwen/Qwen3.6-35B-A3B | EU | 262K | $0.20 | $1.25 |
| sference/deepseek-ai/DeepSeek-V4-Flash-0731 | EU | 1.0M | $0.14 | $0.28 |
| sference/deepseek-v4-flash | EU | 1.0M | $0.14 | $0.28 |
| sference/deepseek-v4-flash-latest | EU | 1.0M | $0.14 | $0.28 |
| sference/kimi-k3 | EU | 1.0M | $2.25 | $11.25 |
| sference/thinkingcap-qwen3.6-27b | EU | 262K | $0.40 | $2.60 |
| sference/zai-org/GLM-5.2 | EU | 1.0M | $1.20 | $4.20 |
Get started
Paste this into your coding agent (Claude Code, Cursor, Codex, and more) and it will set up Opper and route to Sference for you:
Use curl to download, read and follow: https://skills.opper.ai
Then set up Opper to use Sference as the provider, e.g. sference/deepseek-v4-flash.
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:
import OpenAI from "openai";const client = new OpenAI({apiKey: process.env.OPPER_API_KEY,baseURL: "https://api.opper.ai/v3/compat",});const completion = await client.chat.completions.create({model: "sference/deepseek-v4-flash",messages: [{ role: "user", content: "Hello" }],});console.log(completion.choices[0].message.content);
Follow the quick start in our docs for evaluations, fallbacks, and structured output.
About Sference
Sference is a managed inference platform for open models, serving open-weight models, customer fine-tunes, and distillations through one OpenAI-compatible API across realtime, flex, async, and batch workloads. Inference runs on European infrastructure, every model is pinned to a version, zero data retention is the default, and there is no training on customer data.
About Opper AI
Opper AI is the European AI gateway and control plane for agents: one EU-hosted, GDPR-compliant API across 300+ models, with smart routing, automatic fallbacks, built-in evaluations and observability, real-time guardrails, and full OpenAI SDK compatibility.