Ling 3.1 Flash
Ling 3.1 Flash is a hybrid reasoning Mixture-of-Experts model from inclusionAI, with 25B parameters active out of 560B total. It is text in, text out, supports up to 235,929 tokens of output on a 262,144-token context window, and costs $0.075 per million input tokens.
Technical Specifications
Live data fetched from the OpenRouter model registry — always up to date.
What is inclusionai: Ling 3.1 Flash?
Ling 3.1 Flash is a hybrid reasoning model from inclusionAI, a sparse Mixture-of-Experts design with 25B active parameters out of 560B total. It is text-only, which keeps it focused and cheap: $0.075 per million input tokens. What stands out is the output allowance - up to 235,929 tokens, close to the size of its own context window.
Features & Capabilities
Hybrid Reasoning
Reasoning mode plus direct generation in one model.
560B MoE, 25B Active
Sparse design keeps inference cost low at scale.
262K Context
A long document set in a single session.
Deep Output
Up to 235,929 output tokens - high for this price bracket.
Agent Friendly
Suited to pipelines that generate large structured results.
$0.075 per Million
Among the cheapest models that supports long output.
Who Is inclusionai: Ling 3.1 Flash For?
Long-Form Generation
Producing very large outputs without truncation.
Document Processing
Reading a large corpus and returning a full result.
Background Pipelines
Cheap reasoning that runs continuously.
Structured Extraction
Turning long source material into complete records.
inclusionai: Ling 3.1 Flash Pricing
- $0.075 per million input tokens, $0.225 per million output tokens
- Up to 235,929 output tokens - unusual at this price
- Run it from your Hoodgen free credits, no credit card required
- Higher allowances on Hoodgen AI paid plans
What Makes inclusionai: Ling 3.1 Flash Different?
The output allowance is the standout number. Most cheap models cap output somewhere between 8,000 and 32,000 tokens; Ling 3.1 Flash allows up to 235,929, which is nearly the size of its entire context window. For work that produces rather than answers - a full extraction over a long corpus, a large generated document - that removes the chunking and stitching that normally dominates the pipeline.
The architecture explains the price. 25B active parameters out of 560B total is a textbook sparse MoE, which keeps compute per token low while the model still draws on a large pool of knowledge. Hybrid reasoning means it can reason step by step when the task needs it and generate directly when it does not.
It is text-only, so it will not help with images, audio or video. That is the tradeoff for the cost, and for document and code workloads the narrower scope is not a limitation. At $0.075 per million input tokens with that output ceiling, it is a very particular tool that almost nothing else covers.
inclusionai: Ling 3.1 Flash Performance & Pricing
Ling 3.1 Flash competes on output size per rupee spent rather than on multimodal breadth.
What Can You Do With inclusionai: Ling 3.1 Flash?
Ling 3.1 Flash is for text work where the output is large. If a pipeline keeps hitting truncation limits on a cheap model, this is the model to try before reaching for a frontier tier.
Best Use Cases
- Generating complete large documents
- Full-corpus extraction without chunking
- Long-form content pipelines
- Batch analysis that runs continuously
- Anywhere output truncation is the bottleneck
inclusionai: Ling 3.1 Flash Privacy & Data Policy
Ling 3.1 Flash is developed by inclusionAI and served through OpenRouter. Prompts and completions are processed under the provider terms.
As with any third-party model, avoid sending passwords or personal data. Hoodgen AI keeps models with different strengths on the same account so a task can be rerouted if the fit is wrong.
inclusionai: Ling 3.1 Flash vs Other Models
Frequently Asked Questions
Large outputs, tiny price
Chat with Ling 3.1 Flash on Hoodgen - sign in and start in seconds.
Chat with inclusionai: Ling 3.1 Flash