TrendyolLanguage model
Trendyol Asure 12B
Trendyol Asure 12B has no overall index value yet because there are too few independent benchmark results.
23out of 100same as last month
based on two data groups, mostly usage
Summary
Trendyol Asure 12B is a multimodal instruct model built on Gemma 3 12B. It accepts text and images and produces structured answers for e-commerce work in Turkish and English.
The model card says Asure’s general world knowledge is limited, and outputs can be inaccurate or biased. Asure generates text; the card documents no tool-calling interface. The weights are subject to the Gemma Terms of Use and can run locally with Transformers, vLLM, or SGLang. Its documented focus remains structured e-commerce work in Turkish and English.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 3 published benchmark results
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.
TurkishMMLU (Yüksel) (Show measurement, test conditions, and source)
- Source value
- 62.22
- Score
- 62.22
- Metric
- macro-average accuracy across nine subjects
- Unit
- %
- Benchmark version
- turkishmmlu-yuksel-2026.06; EleutherAI lm-evaluation-harness 0.4.12
- Direction
- Higher is better
- Sample
- 900
- Source type
- independent-evaluator
- Evaluator
- YerelAI direct evaluation result artifact
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Independent evaluation. Exact Trendyol/Trendyol-LLM-Asure-12B model SHA df1164d37f47080ba5fbc1ff4defa0c7a1a218a5. lm-evaluation-harness 0.4.12, turkishmmlu, 5-shot, seed 42, bf16, RTX 4090, context length 4096; 9-subject macro average. Result artifact SHA-256 e7bcbd77c279b38b3e3ee4a473f9851cc432716b2674df6661009fe6d23a01c4.
- Context
- This is an independent direct run on the Turkish variant, not a general MMLU score. The structured result pins a model content SHA even though the repository revision label is main.
Belebele (Turkish) (Show measurement, test conditions, and source)
- Source value
- 86.56
- Score
- 86.56
- Metric
- accuracy on the tur_Latn subset
- Unit
- %
- Benchmark version
- belebele-tur-Latn-2026.06; EleutherAI lm-evaluation-harness 0.4.12
- Direction
- Higher is better
- Sample
- 900
- Source type
- independent-evaluator
- Evaluator
- YerelAI direct evaluation result artifact
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Independent evaluation. Exact Trendyol/Trendyol-LLM-Asure-12B model SHA df1164d37f47080ba5fbc1ff4defa0c7a1a218a5. lm-evaluation-harness 0.4.12, task belebele_tur_Latn, 5-shot, apply_chat_template, seed 42, bf16, RTX 5090, context length 4096; 900 items. Result artifact SHA-256 f19b9d98a7880922876baa931fcdd8845c547b40324608ccd517ead2886c5283.
- Context
- This is an independent direct run on the Turkish tur_Latn subset, not an all-language Belebele aggregate. It is a separate run from TurkishMMLU on different hardware.
TurkishMMLU (ort.) (Show measurement, test conditions, and source)
- Source value
- 60.9
- Score
- 60.9
- Metric
- acc
- Unit
- %
- Direction
- Higher is better
- Sample
- 900
- Source type
- independent-evaluator
- Evaluator
- Erk TurkishMMLU evaluation report (ecloudtechnology)
- Status
- active
- Retrieved at
- 2026-10-23
- Methodology
- Independent evaluation. The table identifies Trendyol Asure at 12B. The report says all listed models were evaluated under identical conditions, with the same tool and test set; its method text identifies the full TurkishMMLU test set as 900 questions. This row is the full-set zero-shot result without option-order rotation, using lm-evaluation-harness and accuracy. The separately described contamination-filtered 652-question, option-rotated protocol applies to Erk-14B only, not to this Asure row. The report is published by ecloudtechnology's Erk project. The exact Asure checkpoint, revision, and serving route are not stated.
- Context
- sampleCount 900 is the published size of the full shared test set, not a number of reruns. The report does not provide a cell-specific response-coverage count, an immutable Asure checkpoint/revision, serving route, or confidence interval. Do not substitute Erk-14B's separate clean 652-question result or compare this zero-shot/no-rotation row directly with the stored 5-shot Yüksel run.
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
The 100 most popular models of this category, 0 to 100, higher is more popular
20 of 101 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 7, 2026
Trendyol Asure 12BSonar 23Reka Edge 23Index-NativeLong-2B 23Virtuoso Large 23
Method changes in the timeline, a jump at these points is no market event:
- 1November 1, 2026: Value only with search or usage data, 30-day attention
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Similar models
More models of the same kind as Trendyol Asure 12B, chosen by model line, developer, size, price and release date.
All specifications
Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
Frequently asked questions about Trendyol Asure 12B
How large is the context window of Trendyol Asure 12B?
131,072 tokens. That is how much text Trendyol Asure 12B reads at once, about 200 A4 pages.
Page count as a rule of thumb from Artificial Analysis
Can I run Trendyol Asure 12B locally?
Yes, the weights are open.
Stated by Trendyol
- Open weights
- yes
- License
- Gemma Terms of Use
- Parameters
- 12B
- active per token
- n/a
Sources and data date
Show all 17 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
No active result from the reviewed external evaluators can currently be attributed to this model. Other source evidence remains available below.
Without vendor claims and historical or disputed results; those stay labelled in the single values.
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved October 27, 2026. |
| Output | Documented modalities. Retrieved October 27, 2026. |
| Research status | Research date October 15, 2026. 8 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not published | additional-official-numeric-benchmarks Note The exact official card revision adds no new numeric evaluation table. Previously recorded YerelAI rows are not copied or relabeled as official scores. |
| Additional source | Hugging Face (retrieved October 15, 2026) · Trendyol Asure 12B model card · Editorial description reviewed October 2, 2026 |
| Additional source | YerelAI direct evaluation result artifact (retrieved October 15, 2026)12 |
| Additional source | Erk TurkishMMLU evaluation report (ecloudtechnology) (retrieved October 23, 2026) |
| Additional source | Model metadata (retrieved October 15, 2026)12 |
| Additional source | Model metadata (retrieved October 16, 2026) |
| Additional source | Model metadata (retrieved October 23, 2026) |
| Additional source | Trendyol Asure 12B chat template · Editorial description reviewed October 2, 2026 |
| Additional source | Gemma Terms of Use · Editorial description reviewed October 2, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Ollama: cumulative downloads for the popularity index |
| Additional source | Replicate: cumulative runs for the popularity index |
| Additional source | Google Trending Now: limited feed selection for Germany and the US |
| Additional source | Creative Commons Attribution 4.0 |