Ling 3.1 Flash 1019 Benchmarks, Specifications & Availability

Ling 3.1 Flash 1019

Explore Ling 3.1 Flash 1019 from inclusionAI: published specifications, source-linked vendor benchmarks, independent evaluator coverage and recorded pricing when available.

Latest versionPrevious version: Ling 3.1 FlashRelease source ↗

Published specifications

Provider: inclusionAI
Access: Open weights
License: MIT
Context window: 1M
Total parameters: 560B
Active parameters: 25B
Released: 2026-10-27
Modalities: text
Family: Ling 3.1

Model card & announcement

Model notes

Dated post-training upgrade of Ling-3.1-flash, separate from the September checkpoint. Official ungated MIT weights; repository created 27 October 05:46 UTC, weights uploaded 10:45 UTC. Hybrid-linear MoE, 560B total / 25B active per the card (Hugging Face safetensors metadata counts 572,120,680,768 parameters). The card states a 1M-token context and Artificial Analysis lists 1,048,576 tokens; the shipped config.json sets max_position_embeddings 262,144 with no rope_scaling, so a stock load without a serving-side extension serves 256K. The card links only the generic inclusionai/ling-3.1-flash OpenRouter route; no dated route or endpoint evidence exists, so no price is attached. Card AA figures are vendor citations of Artificial Analysis Intelligence Index v4.3.2; the independent AA page links these exact weights.

Compare Ling 3.1 Flash 1019 with other models →Compare subscription plans Explore data coverage

Pricing · OpenRouter

No OpenRouter pricing is recorded for this model. Missing rates are not free.

Official / vendor benchmarks

Default headline records. Own-vendor, peer-vendor and third-party provenance remain visible in evidence; configurations may differ.

Official / vendor headline scores; expand evidence for every record
Benchmark / evaluator	Headline score	Evidence
AA Intelligence Index v4.3.2 (vendor-cited)	44unknown effort	All 1 recorded result & sources 44 · raw 44 index Headline · unknown effort · Own vendor Source/record date: 2026-10-27 Harness: Artificial Analysis evaluation as cited by InclusionAI on the Ling-3.1-flash-1019 card (introduction text, Intelligence Index v4.3.2 context); vendor citation of a third-party evaluator, separate from independent AA observations. Effort label not published; card recommends thinking enabled. Moved 2026-10-28 from the mixed-version vendor-cited column to the v4.3.2 column; the source states v4.3.2. https://huggingface.co/inclusionAI/Ling-3.1-flash-1019/blob/dddfee4bb43519a92761fbc5cefe3f0d3ff4aded/README.md
AutomationBench-AA (vendor-reported)	68%unknown effort	All 1 recorded result & sources 68% · raw 68 % Headline · unknown effort · Own vendor Source/record date: 2026-10-27 Harness: Artificial Analysis evaluation as cited by InclusionAI on the Ling-3.1-flash-1019 card (introduction text, Intelligence Index v4.3.2 context); vendor citation of a third-party evaluator, separate from independent AA observations. Effort label not published; card recommends thinking enabled. https://huggingface.co/inclusionAI/Ling-3.1-flash-1019/blob/dddfee4bb43519a92761fbc5cefe3f0d3ff4aded/README.md
CritPt	26%unknown effort	All 1 recorded result & sources 26% · raw 26 % Headline · unknown effort · Own vendor Source/record date: 2026-10-27 Harness: Artificial Analysis evaluation as cited by InclusionAI on the Ling-3.1-flash-1019 card (introduction text, Intelligence Index v4.3.2 context); vendor citation of a third-party evaluator, separate from independent AA observations. Effort label not published; card recommends thinking enabled. https://huggingface.co/inclusionAI/Ling-3.1-flash-1019/blob/dddfee4bb43519a92761fbc5cefe3f0d3ff4aded/README.md
SciCode	57%unknown effort	All 1 recorded result & sources 57% · raw 57 % Headline · unknown effort · Own vendor Source/record date: 2026-10-27 Harness: Artificial Analysis evaluation as cited by InclusionAI on the Ling-3.1-flash-1019 card (introduction text, Intelligence Index v4.3.2 context); vendor citation of a third-party evaluator, separate from independent AA observations. Effort label not published; card recommends thinking enabled. https://huggingface.co/inclusionAI/Ling-3.1-flash-1019/blob/dddfee4bb43519a92761fbc5cefe3f0d3ff4aded/README.md
AA-Briefcase (vendor-reported)	1524unknown effort	All 1 recorded result & sources 1524 · raw 1524 Elo Headline · unknown effort · Own vendor Source/record date: 2026-10-27 Harness: Artificial Analysis evaluation as cited by InclusionAI on the Ling-3.1-flash-1019 card (introduction text, Intelligence Index v4.3.2 context); vendor citation of a third-party evaluator, separate from independent AA observations. Effort label not published; card recommends thinking enabled. https://huggingface.co/inclusionAI/Ling-3.1-flash-1019/blob/dddfee4bb43519a92761fbc5cefe3f0d3ff4aded/README.md

Independent evaluators

Evaluator harnesses are distinct from vendor measurements. Missing coverage is not a failed test.

Independent evaluator headline scores; expand evidence for every record
Benchmark / evaluator	Headline score	Evidence
Intelligence Index v4.3.2 · Artificial Analysis	44unknown effort	All 1 recorded result & sources 44 · raw 44 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-25 Manual public rendered leaderboard observation on 25 October 2026 (research-raw/kimi-raon-2026-10-25/evaluators/aa-full-ax.txt), while Intelligence Index v4.3.2 was current; the AA changelog of 22 October also lists Intelligence Index 44. The exact AA model page links the official inclusionAI/Ling-3.1-flash-1019 weights (checked 28 October). AA marks the model as reasoning but publishes no effort label. Since 27 October AA shows Intelligence Index v5.0 values, which are not recorded here. No API, internal app state, scheduled import, hidden precision or inferred run date. Version 4.3.2 ten-component index, not task accuracy. https://artificialanalysis.ai/leaderboards/models
Cost per Intelligence Index task · Artificial Analysis	$1.49/taskunknown effort	All 1 recorded result & sources $1.49/task · raw 1.49 USD/task Headline · unknown effort · Independent evaluator Source/record date: 2026-10-25 Manual public rendered leaderboard observation on 25 October 2026 (research-raw/kimi-raon-2026-10-25/evaluators/aa-full-ax.txt), while Intelligence Index v4.3.2 was current; the AA changelog of 22 October also lists Intelligence Index 44. The exact AA model page links the official inclusionAI/Ling-3.1-flash-1019 weights (checked 28 October). AA marks the model as reasoning but publishes no effort label. Since 27 October AA shows Intelligence Index v5.0 values, which are not recorded here. No API, internal app state, scheduled import, hidden precision or inferred run date. Weighted workload-specific API-equivalent cost, not a token tariff. https://artificialanalysis.ai/leaderboards/models

Read how we select and source scores or the comparison guide.