MiMo V2.6 Flash on Novita AI: Model Launch and Pricing

MiMo V2.6 Flash is available on Novita AI as a serverless reasoning model for teams that need multimodal input, a large context window, and lower per-token pricing for frequent calls. The live Novita catalog lists the model as xiaomimimo/mimo-v2.6-flash, with text, image, video, and audio input; a 1,048,576-token context window; up to 131,072 output tokens; and current prices of $0.14 per 1M input tokens, $0.0028 per 1M cache-read input tokens, and $0.28 per 1M output tokens. For a live configuration and pricing check, open the MiMo V2.6 Flash model page.

MiMo V2.6 Flash at a glance

Field	Current Novita AI listing
Display name	MiMo V2.6 Flash
Model ID	`xiaomimimo/mimo-v2.6-flash`
Access	Serverless API
Input modalities	Text, image, video, and audio
Output modality	Text
Context window	1,048,576 tokens (1M)
Maximum output	131,072 tokens (128K)
Listed features	Reasoning, function calling, structured outputs
Endpoint families	Chat Completions, Responses, and Anthropic-compatible
Input price	$0.14 per 1M tokens
Cache-read input price	$0.0028 per 1M tokens
Output price	$0.28 per 1M tokens

The model’s positioning is clear: it is designed for high-frequency calls and large-scale professional workflows where cost matters alongside long-context and multimodal understanding. That makes it a useful model to test for workloads such as document-heavy agents, media-aware support flows, batch extraction, and tool-using applications that need structured text responses.

What MiMo V2.6 Flash is for

MiMo V2.6 Flash is the cost-efficient variant in Xiaomi’s MiMo V2.6 line. On Novita AI, the hosted model combines four input types—text, image, video, and audio—with text output. It also exposes reasoning, function calling, structured outputs, streaming, and prompt-caching capabilities in the live model catalog.

That combination gives developers a flexible starting point, but it does not make the model a default choice for every workload. A team processing short, text-only requests may prefer a smaller or more specialized endpoint after testing. A team that needs a stable conversation policy, source-grounded answers, or strict JSON contracts should test those behaviors on representative data before moving traffic.

The practical advantage of the Flash variant is its price profile. At the current listed rates, it is a sensible candidate when an application has many requests, long reusable instructions, or large volumes of multimodal material to classify and summarize. Cost still depends on actual prompt length, output length, caching behavior, and retries, so the listed token rates should be treated as planning inputs rather than a billing guarantee.

Availability and API details on Novita AI

Novita AI lists MiMo V2.6 Flash as a serverless model with the exact ID xiaomimimo/mimo-v2.6-flash. The catalog reports chat/completions, responses, and anthropic endpoint families, so applications can select the interface that best fits their existing client and integration pattern.

For OpenAI-compatible integrations, use Novita AI’s documented API surface and set the model field to the exact ID shown above. The current Create chat completion API reference is the right place to confirm authentication, request fields, supported message formats, and response handling before implementation.

Start with a narrow evaluation request instead of sending a production-sized context immediately. For example, test a text-only task first, then add images, audio, or video one at a time. This makes it easier to isolate model-quality questions from payload formatting, latency, upload, or application-side parsing issues.

Pricing and context limits

The following prices and limits were checked against the Novita AI model catalog on September 30, 2026:

Input tokens: $0.14 per 1M tokens
Cache-read input tokens: $0.0028 per 1M tokens
Output tokens: $0.28 per 1M tokens
Context window: 1,048,576 tokens
Maximum output: 131,072 tokens

The cache-read price is particularly relevant for applications that reuse a stable system prompt, policy, product catalog, or long reference context. It is not an automatic discount on every request: your application still needs to follow the current platform behavior for prompt caching, and the token classification should be verified in your account before it drives an architecture decision.

The 1M context figure describes the catalog’s maximum context capacity, not a recommendation to fill every request. Very large prompts can increase cost, require more application-side retrieval discipline, and make failures harder to diagnose. In most production systems, retrieval, chunking, and output caps remain valuable even when the model accepts a large context.

When MiMo V2.6 Flash is a good fit

Choose MiMo V2.6 Flash for an evaluation when your workload benefits from several of the following:

High request volume: The current input and output rates are suited to teams comparing cost-sensitive serverless options.
Mixed media inputs: The catalog lists image, video, and audio alongside text, so you can test a single model path for multimodal intake.
Long working context: A 1M-token context window can help when a workflow needs to inspect a large collection of retrieved material, transcripts, or agent state.
Tool-driven workflows: Function calling and structured-output support make it relevant for applications that need machine-readable model responses.
Repeated reference context: The listed cache-read rate is worth evaluating for workloads with reusable prompts or knowledge blocks.

It is less likely to be the first choice when you need a published benchmark result for a specific domain, a particular response format that has not been tested in your environment, or a model with a narrower specialized capability. The public Novita catalog provides capability and pricing metadata, not a workload-specific quality guarantee. Run a side-by-side evaluation using your own prompts, target languages, media samples, and failure cases.

How to evaluate it safely

An effective first evaluation can be small and deliberate:

Pick one measurable task. Use a real classification, extraction, or support workflow with expected outputs, rather than a collection of unrelated demo prompts.
Set a practical output cap. The model can return up to 128K tokens, but shorter caps make latency, cost, and error handling easier to inspect.
Test modalities separately. Establish a text baseline before adding image, audio, or video inputs, then compare the result quality and operational cost.
Validate structured responses. When a workflow uses JSON or function calls, validate the returned data in the application and define a retry or repair path.
Measure cache behavior. Repeat representative requests with stable context and compare usage records before assuming cache-read economics in production.

For teams building an agent that needs a lighter long-context model, Ling-3.0 Tiny on Novita AI is another useful point of comparison. For a different Flash-model profile with multimodal input, see Qwen3.8 Flash on Novita AI.

Conclusion

MiMo V2.6 Flash gives Novita AI developers a serverless option for multimodal, long-context, and tool-using workflows with an efficiency-oriented price profile. The live listing currently combines a 1M-token context window, up to 128K output, text/image/video/audio input, and $0.14/$0.0028/$0.28 pricing per 1M input/cache-read/output tokens.

Start with a representative task and compare the result, token usage, and operational behavior against the alternatives you already run. If the model’s multimodal support, context capacity, and current rates line up with your workload, explore MiMo V2.6 Flash on Novita AI before making it part of a production routing policy.

FAQ

What is the MiMo V2.6 Flash model ID on Novita AI?

Use xiaomimimo/mimo-v2.6-flash.

What inputs does MiMo V2.6 Flash support?

The current Novita AI catalog lists text, image, video, and audio inputs, with text output.

What is the context window for MiMo V2.6 Flash?

The current catalog lists a 1,048,576-token context window and up to 131,072 output tokens.

What are the current MiMo V2.6 Flash prices on Novita AI?

As checked on September 30, 2026, Novita AI lists $0.14 per 1M input tokens, $0.0028 per 1M cache-read input tokens, and $0.28 per 1M output tokens. Recheck the live model page before production budgeting.