Skip to main content
AI Choice Engine

InclusionAI

Ant Ling 3.0 Flash

Ant Ling 3.0 Flash is a InclusionAI model with 262K context, $0.021 in / $0.063 out per million tokens, last verified October 25, 2026. Open weights: yes. Public-board scores are linked to the publisher; vendor-only figures stay in the watch-out, not on the board.

Ling 3.0 Flash is Ant Group's open-weight (MIT) 124B/5.1B agentic MoE from a fintech lab.

Recorded profile verification October 25, 2026Individual provider fields may change

AI model specification details
SpecificationAnt Ling 3.0 Flash
ProviderInclusionAI
TierBudget
Context window262K
Max output33K
Input / 1M tokens$0.021
Output / 1M tokens$0.063
WeightsOpen
Parameters124B total / 5.1B active (MoE)
Reasoning levelsNot verifiedUnverified
Modalitiestext
LicenseMIT (publisher repository metadata)
API model idinclusionai/ling-3.0-flash
ReleasedJuly 23, 2026

Pricing tiers: Novita via the current OpenRouter endpoint feed: $0.021/$0.063 per MTok, cache read $0.0042, 65% route discount with no published end date, 262,144 context and 32,768 maximum completion. DeepInfra is $0.06/$0.18, cache $0.012, with 131,072 context and the same output cap. Ant-direct native billing is CNY: current 72% promotional ¥0.14/¥0.42, cache ¥0.028; regular ¥0.50/¥1.50, cache ¥0.10. Ant's third-party summary quotes lower Flash USD rates than the captured live OpenRouter endpoints and explicitly defers to platform pricing; the endpoint-scoped rates are used here. These tariffs do not establish direct Ant USD billing or a free entitlement.

Publisher evidence

Published benchmark results

Each result keeps its source date visible. Dates refer to source publication or capture unless the source explicitly states when it measured the result. A missing benchmark is not treated as a zero.

Published benchmark results for Ant Ling 3.0 Flash
BenchmarkScoreSource dateSource
Artificial Analysis Intelligence IndexLing 3.0 Flash202026-10-01Artificial Analysis · View source

Best for

  • Open-weight deployments
  • Agentic work
  • Cost-sensitive

Watch out

Native config gives 262,144 positions; the article describes extension to 1M, not the default. Thinking is enabled by default and can be disabled through chat_template_kwargs.enable_thinking=false; no graded effort enum is verified. Recommended sampling is temperature=0.6/top_p=0.95/top_k=20. Published SWE/Terminal results use named harnesses and budgets, not universal output limits. July 23 is the native-platform launch; July 24 is the article date and public weights are a separate stage. First-party ID Ling-3.0-flash is documented; the profile API ID selects OpenRouter. Exact weights release day remains unverified.

History

Ant Ling 3.0 Flash timeline

Release and status events for this model that passed two-source verification — dated, sourced, and linked.

Source receipts

Catalog figures for this model were checked against the following sources.

Common questions

Ant Ling 3.0 Flash

Answered from the verified figures on this page rather than general guidance.

How much does Ant Ling 3.0 Flash cost per million tokens?
Ant Ling 3.0 Flash lists base rates of $0.021 per million input tokens and $0.063 per million output tokens. Estimate both input and output token volume and any additional charges for your workload. Novita via the current OpenRouter endpoint feed: $0.021/$0.063 per MTok, cache read $0.0042, 65% route discount with no published end date, 262,144 context and 32,768 maximum completion. DeepInfra is $0.06/$0.18, cache $0.012, with 131,072 context and the same output cap. Ant-direct native billing is CNY: current 72% promotional ¥0.14/¥0.42, cache ¥0.028; regular ¥0.50/¥1.50, cache ¥0.10. Ant's third-party summary quotes lower Flash USD rates than the captured live OpenRouter endpoints and explicitly defers to platform pricing; the endpoint-scoped rates are used here. These tariffs do not establish direct Ant USD billing or a free entitlement.
What is Ant Ling 3.0 Flash's context window?
Ant Ling 3.0 Flash accepts about 262K tokens of context. That only matters if you routinely send very long documents, large codebases, or multi-turn histories that approach that limit.
What is Ant Ling 3.0 Flash best for?
Ant Ling 3.0 Flash is a budget tier from InclusionAI. It suits open-weight deployments, agentic work, cost-sensitive. Native config gives 262,144 positions; the article describes extension to 1M, not the default. Thinking is enabled by default and can be disabled through chat_template_kwargs.enable_thinking=false; no graded effort enum is verified. Recommended sampling is temperature=0.6/top_p=0.95/top_k=20. Published SWE/Terminal results use named harnesses and budgets, not universal output limits. July 23 is the native-platform launch; July 24 is the article date and public weights are a separate stage. First-party ID Ling-3.0-flash is documented; the profile API ID selects OpenRouter. Exact weights release day remains unverified.
Can I self-host Ant Ling 3.0 Flash?
Ant Ling 3.0 Flash publishes open weights, but self-hosting depends on the licence, hardware footprint, quantisation quality, and serving stack. A hosted API is often cheaper until you have measured throughput and concurrency on your own hardware.
Model comparison catalogBenchmark library