Language modelOpen weights

LLaDA 2.2 Mini

inclusionAI

Released: -
Data date: October 7, 2026

Input: Text
Output: Text

LLaDA 2.2 Mini uses Levenshtein editing to modify sequences during parallel generation through DELETE and INSERT control tokens. The MoE diffusion model has 16 billion parameters excluding embeddings, activates 1.4 billion, and supports a provider-reported 128K context.

Block Routing limits expert activation at the diffusion-block level, targeting agentic tasks with long contexts. The Apache 2.0 weights are documented for local deployment; Hugging Face currently lists no hosted inference provider.

Specifications and access

Specification	Value and source
Provider model ID	inclusionAI/LLaDA2.2-miniSource
Access	Open weights for local deploymentSource
Input modalities	TextSource
Output modalities	TextSource
License	Apache 2.0Source
Reported context	128K tokens; no exact integer limit statedSource
Parameter count scope	16B total parameters excluding embeddings, with 1.4B activated during inference.Source

Measurements without matching peer values

These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.

BFCL v4 (Show measurement, test conditions, and source)

Source value: 47.68
Score: 47.68
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Benchmark version: v4
Category: tool-use
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

BFCL v3 (Show measurement, test conditions, and source)

Source value: 69.02
Score: 69.02
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Benchmark version: v3
Category: tool-use
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

AIME 2026 (Show measurement, test conditions, and source)

Source value: 35.05
Score: 35.05
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: math
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

OlympiadBench (Show measurement, test conditions, and source)

Source value: 61.11
Score: 61.11
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: math
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

LiveCodeBench v6 (Show measurement, test conditions, and source)

Source value: 28.14
Score: 28.14
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Benchmark version: v6
Category: coding
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

MultiPL-E (Show measurement, test conditions, and source)

Source value: 65.26
Score: 65.26
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: coding
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

IFBench (Show measurement, test conditions, and source)

Source value: 24.93
Score: 24.93
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: instruction-following
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

Multi-IF (Show measurement, test conditions, and source)

Source value: 57.03
Score: 57.03
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: instruction-following
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

KOR-Bench (Show measurement, test conditions, and source)

Source value: 43.6
Score: 43.6
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: reasoning
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

GPQA-Diamond (Show measurement, test conditions, and source)

Source value: 44.41
Score: 44.41
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: knowledge
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

LongBench v2 (Show measurement, test conditions, and source)

Source value: 34.99
Score: 34.99
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Benchmark version: v2
Category: long-context
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

General Average (Show measurement, test conditions, and source)

Source value: 46.47
Score: 46.47
Metric: provider-reported score; scale/unit not stated in table
Unit: No unit provided
Category: aggregate
Direction: Higher is better
Source type: vendor-reported
Evaluator: inclusionAI LLaDA2.2-mini model card
Status: active
Retrieved at: 2026-10-07
Methodology: Vendor-reported result. Provider comparison-table value from the exact LLaDA2.2-mini column. The card does not state evaluation harness, sampling settings, benchmark revision or sample counts; scores are preserved as shown without assigning a unit.
Context: Provider-reported table values. Results for LLaDA2.0-mini and LLaDA2.1-mini are not included or transferred to LLaDA2.2-mini.

Source inclusionAI LLaDA2.2-mini model card

Page 1 of 2

Sources and data date

Type	Evidence and data date
Input	Documented modalities. Retrieved October 7, 2026. Source: LLaDA 2.2 Mini model card · LLaDA 2.2 Mini model card (retrieved October 7, 2026) · LLaDA 2.2 Mini model card · Editorial description reviewed October 7, 2026
Output	Documented modalities. Retrieved October 7, 2026. Source: LLaDA 2.2 Mini model card · LLaDA 2.2 Mini model card (retrieved October 7, 2026) · LLaDA 2.2 Mini model card · Editorial description reviewed October 7, 2026
Research status	Research date October 7, 2026. 2 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Unresolved	evaluation-protocol-and-independent-results Original wording The model card publishes no harness, sampling, version or sample-count details for its table. No independent evaluation of the exact LLaDA2.2-mini was verified in this pass.