Opper AI partners with Munito AI for private open-model inference in Europe
By Felix Wunderlich -
Stockholm, Sweden, November 2026. Opper AI and Munito AI are partnering to bring Munito's open-model inference to the Opper AI gateway. Builders can now reach Munito's catalogue through the Opper API, from DeepSeek V4 Flash, DeepSeek V4.1 Flash and GLM-5.3-Flash to Qwen3.8 27B and Aleph Alpha's Kolibri 1.
Why this matters
Built by the engineer who led inference at Mistral AI. Munito AI is a Paris company backed by Y Combinator, founded by Alexey Eryshev, whose team at Mistral AI built its API and inference products. Munito serves open-weight models from DeepSeek, Z.ai, Qwen and Aleph Alpha, including Kolibri 1, Aleph Alpha's open English-German model for regulated work, with its full 1M-token context.
Pick the price for each call. Every Munito model comes in three tiers, and on Opper the tier is part of the model string. Flex (munito:flex/…) costs about half the standard price for batch jobs and evaluations that can wait, standard (munito/…) is the everyday call, and priority (munito:priority/…) puts interactive traffic first. At launch its flex tier is the cheapest route on the gateway, in any region, for Kolibri 1 and GLM-5.3-Flash, and the cheapest EU route for DeepSeek V4.1 Flash. Prices move, so the table further down is fetched live, and our flex vs standard test shows what each tier means for speed.
Clear data handling. Munito AI is a French company that serves its models from three hosts: the EU host runs inference in EU member states, the US host in the United States, and the global host in any region. On every host, prompts and completions are not logged, held for abuse monitoring or used for training, and a DPA is available. Full residency and retention detail per route lives on the Munito AI provider page.
More than a router: Rules for every call. Routing to Munito is just the entry point. Every call through Opper can run under your Rules: spend limits, data retention, model access, content checks and routing order, set once for the organization or per project. Pin Munito's flex tier for batch work or set Munito as a fallback, and because Opper is OpenAI-compatible, getting there is a model string, not a migration.
"Open models pay off when they run privately and at a price that fits the job. Munito brings inference know-how from Mistral AI and three tiers on every model, so builders on Opper can choose speed or savings call by call."
— Göran Sandahl, Co-founder and CEO, Opper AI
"We started Munito so European companies can run the best open models, private by default, without managing a single GPU. Opper puts our models in front of a wide community of builders through one API, next to every other model they already use."
— Alexey Eryshev, Founder and CEO, Munito AI
Models live today
The catalog below is fetched live from Opper's model API and filtered to Munito-hosted models, grouped by host with each model's flex, standard and priority tiers listed together. Availability, context windows, and pricing stay in sync with what's actually callable through Opper. The region column shows where each route runs.
| Model | Region | Context | Input / 1M | Output / 1M |
|---|---|---|---|---|
| munito:flex/deepseek-v4-flash | EU | 1.0M | $0.06 | $0.14 |
| munito/deepseek-v4-flash | EU | 1.0M | $0.11 | $0.28 |
| munito:priority/deepseek-v4-flash | EU | 1.0M | $0.19 | $0.49 |
| munito:flex/deepseek-v4.1-flash-1017 | EU | 1.0M | $0.15 | $0.60 |
| munito/deepseek-v4.1-flash-1017 | EU | 1.0M | $0.30 | $1.20 |
| munito:priority/deepseek-v4.1-flash-1017 | EU | 1.0M | $0.53 | $2.10 |
| munito:flex/glm-5.3-flash | EU | 1.0M | $0.07 | $0.25 |
| munito/glm-5.3-flash | EU | 1.0M | $0.15 | $0.50 |
| munito:priority/glm-5.3-flash | EU | 1.0M | $0.26 | $0.88 |
| munito:flex/kolibri-1 | EU | 1.0M | $0.08 | $0.28 |
| munito/kolibri-1 | EU | 1.0M | $0.16 | $0.56 |
| munito:priority/kolibri-1 | EU | 1.0M | $0.28 | $0.98 |
| munito:flex/qwen3.8-27b | EU | 262K | $0.10 | $1.40 |
| munito/qwen3.8-27b | EU | 262K | $0.20 | $2.70 |
| munito:priority/qwen3.8-27b | EU | 262K | $0.35 | $4.70 |
| munito:global:flex/deepseek-v4-flash | GLOBAL | 1.0M | $0.05 | $0.13 |
| munito:global/deepseek-v4-flash | GLOBAL | 1.0M | $0.10 | $0.25 |
| munito:global:priority/deepseek-v4-flash | GLOBAL | 1.0M | $0.17 | $0.44 |
| munito:global:flex/deepseek-v4.1-flash-1017 | GLOBAL | 1.0M | $0.14 | $0.50 |
| munito:global/deepseek-v4.1-flash-1017 | GLOBAL | 1.0M | $0.27 | $1.00 |
| munito:global:priority/deepseek-v4.1-flash-1017 | GLOBAL | 1.0M | $0.47 | $1.80 |
| munito:global:flex/glm-5.3-flash | GLOBAL | 1.0M | $0.07 | $0.23 |
| munito:global/glm-5.3-flash | GLOBAL | 1.0M | $0.13 | $0.45 |
| munito:global:priority/glm-5.3-flash | GLOBAL | 1.0M | $0.23 | $0.79 |
| munito:global:flex/kolibri-1 | GLOBAL | 1.0M | $0.07 | $0.26 |
| munito:global/kolibri-1 | GLOBAL | 1.0M | $0.13 | $0.51 |
| munito:global:priority/kolibri-1 | GLOBAL | 1.0M | $0.23 | $0.89 |
| munito:global:flex/qwen3.8-27b | GLOBAL | 262K | $0.09 | $1.20 |
| munito:global/qwen3.8-27b | GLOBAL | 262K | $0.18 | $2.40 |
| munito:global:priority/qwen3.8-27b | GLOBAL | 262K | $0.32 | $4.20 |
| munito:us:flex/qwen3.8-27b | US | 262K | $0.10 | $1.40 |
| munito:us/qwen3.8-27b | US | 262K | $0.20 | $2.70 |
| munito:us:priority/qwen3.8-27b | US | 262K | $0.35 | $4.70 |
Get started
Paste this into your coding agent (Claude Code, Cursor, Codex, and more) and it will set up Opper and route to Munito for you:
Use curl to download, read and follow: https://skills.opper.ai
Then set up Opper to use Munito AI as the provider, e.g. munito/glm-5.3-flash, or munito:flex/glm-5.3-flash for batch work.
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: "munito/glm-5.3-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 Munito AI
Munito AI is an open-model AI cloud founded in 2026 and based in Paris, France, and backed by Y Combinator. It runs the GPUs, the inference and sandboxes for agents, in Europe, in a customer's own cloud or on-premises, with zero data retention by default.
About Opper AI
Opper AI is the European AI gateway for agents, based in Stockholm. It gives developers access to 700+ leading AI models through one EU-hosted, GDPR-compliant API, pay-as-you-go, with spend caps, model allowlists, zero data retention and full OpenAI SDK compatibility.