# MultiNRC > Source: Benchize (https://benchize.com) — https://benchize.com/benchmarks/scale%3Amultinrc. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/benchmarks/scale%3Amultinrc. Motivated by this gap, we developed the Multilingual Native Reasoning Challenge (MultiNRC), a new multilingual evaluation benchmark that contains native and challenging reasoning questions of the targeted language, to assess LLMs' reasoning capability under a natural and native context. Specifically, MultiNRC contains four categories of reasoning questions: language-specific linguistic reasoning, wordplay riddles, cultural/tradition reasoning, and finally mathematical reasoning with cultural relevance. We first release MultiNRC in French, Spanish, and Chinese. We present the definitions of the 4 reasoning categories as below. Language-specific Linguistic Reasoningquestions are based on grammatical rules, honorifics, or language-specific conventions that only exist in the target language but not English. Such questions require reasoning about language structure or usage. Examples under this category include word formation problems, relation inference based on language conventions, etc. We enforce that the questions should require multi-step reasoning, instead of mere linguistic feature identification without reasoning. In the French example shown below, the question requires recognizing that the noun délice is masculine in the singular and feminine in the plural, an uncommon grammatical characteristic not present in English. Canonical URL: https://benchize.com/benchmarks/scale%3Amultinrc Last updated: Wed, 21 Oct 2026 18:52:20 GMT Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations. Motivated by this gap, we developed the Multilingual Native Reasoning Challenge (MultiNRC), a new multilingual evaluation benchmark that contains native and challenging reasoning questions of the targeted language, to assess LLMs' reasoning capability under a natural and native context. Specifically, MultiNRC contains four categories of reasoning questions: language-specific linguistic reasoning, wordplay riddles, cultural/tradition reasoning, and finally mathematical reasoning with cultural relevance. We first release MultiNRC in French, Spanish, and Chinese. We present the definitions of the 4 reasoning categories as below. Language-specific Linguistic Reasoningquestions are based on grammatical rules, honorifics, or language-specific conventions that only exist in the target language but not English. Such questions require reasoning about language structure or usage. Examples under this category include word formation problems, relation inference based on language conventions, etc. We enforce that the questions should require multi-step reasoning, instead of mere linguistic feature identification without reasoning. In the French example shown below, the question requires recognizing that the noun délice is masculine in the singular and feminine in the plural, an uncommon grammatical characteristic not present in English. Language & multilingual · Models tested: 15 · Score publishers: [scale](https://labs.scale.com/leaderboard/multinrc) ## Model results | # | Model | Score | ± | | | --- | --- | --- | --- | --- | | 1 | [Muse Spark 1.1](https://benchize.com/models/muse-spark-1-1.md) | 65.59source score | CI 62.72–68.46 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 65.59 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Muse Spark 1.1 sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 62.72–68.46 contaminationWarning Not reported rank 1 reportedConfidenceHalfWidth 2.87 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label Muse Spark 1.1 Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/0` | | | | | | 2 | [Gemini 3.1 Pro Preview](https://benchize.com/models/gemini-3-1-pro-preview.md) | 64.74source score | CI 61.86–67.62 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 64.74 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gemini-3.1-pro-preview sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 61.86–67.62 contaminationWarning Potential contamination warning: This model was evaluated after the public release of MultiNRC, allowing model builder access to the prompts and solutions. rank 3 reportedConfidenceHalfWidth 2.88 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label gemini-3.1-pro-preview Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/2` | | | | | | 3 | [Claude Opus 4.6](https://benchize.com/models/claude-opus-4-6-off.md)Non-thinking | 48.34source score | CI 45.32–51.36 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 48.34 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort non-thinking Model configuration claude-opus-4-6 (Non-Thinking) sourceModelVersion Not reported thinkingEnabled false ### Reported statistics Uncertainty Confidence interval: 45.32–51.36 contaminationWarning Potential contamination warning: This model was evaluated after the public release of MultiNRC, allowing model builder access to the prompts and solutions. rank 13 reportedConfidenceHalfWidth 3.02 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label claude-opus-4-6 (Non-Thinking) Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/12` | | | | | | 4 | [Gemini 2.5 Pro Preview 06-05](https://benchize.com/models/gemini-2-5-pro-preview.md) | 45.12source score | CI 42.12–48.12 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 45.12 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini-2.5-Pro-Preview-06-05 sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 42.12–48.12 contaminationWarning Not reported rank 15 reportedConfidenceHalfWidth 3 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label Gemini-2.5-Pro-Preview-06-05 Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/14` | | | | | | 5 | [R1 0528](https://benchize.com/models/deepseek-r1-0528.md) | 27.58source score | CI 24.88–30.28 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 27.58 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Deepseek-R1-0528 sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 24.88–30.28 contaminationWarning Not reported rank 27 reportedConfidenceHalfWidth 2.7 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label Deepseek-R1-0528 Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/26` | | | | | | 6 | [Qwen3 235B A22B Thinking 2507](https://benchize.com/models/qwen3-235b-a22b-thinking-2507-thinking.md)Thinking | 27.11source score | CI 24.43–29.79 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 27.11 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Qwen3-235B-A22B-Thinking-2507 sourceModelVersion Not reported thinkingEnabled true ### Reported statistics Uncertainty Confidence interval: 24.43–29.79 contaminationWarning Potential contamination warning: This model was evaluated after the public release of MultiNRC, allowing model builder access to the prompts and solutions. rank 28 reportedConfidenceHalfWidth 2.68 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label Qwen3-235B-A22B-Thinking-2507 Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/27` | | | | | | 7 | [Gemini 3.1 Flash Lite Preview](https://benchize.com/models/gemini-3-1-flash-lite-preview.md) | 25.02source score | CI 22.41–27.63 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 25.02 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gemini-3.1-flash-lite-preview sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 22.41–27.63 contaminationWarning Potential contamination warning: This model was evaluated after the public release of MultiNRC, allowing model builder access to the prompts and solutions. rank 29 reportedConfidenceHalfWidth 2.61 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label gemini-3.1-flash-lite-preview Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/28` | | | | | | 8 | [DeepSeek V3.1](https://benchize.com/models/deepseek-v3-1.md) | 23.6source score | CI 21.04–26.16 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 23.6 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration DeepSeek-V3.1 sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 21.04–26.16 contaminationWarning Potential contamination warning: This model was evaluated after the public release of MultiNRC, allowing model builder access to the prompts and solutions. rank 32 reportedConfidenceHalfWidth 2.56 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label DeepSeek-V3.1 Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/31` | | | | | | 9 | [o4 Mini High](https://benchize.com/models/o4-mini-high.md) | 22.18source score | CI 19.67–24.69 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 22.18 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration o4-mini-high sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 19.67–24.69 contaminationWarning Not reported rank 33 reportedConfidenceHalfWidth 2.51 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label o4-mini-high Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/32` | | | | | | 10 | [GPT-4.1](https://benchize.com/models/gpt-4-1.md) | 21.23source score | CI 18.76–23.7 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 21.23 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration GPT-4.1 sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 18.76–23.7 contaminationWarning Not reported rank 34 reportedConfidenceHalfWidth 2.47 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label GPT-4.1 Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/33` | | | | | | 11 | [Qwen3 235B A22B](https://benchize.com/models/qwen3-235b-a22b.md) | 17.63source score | CI 15.33–19.93 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 17.63 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Qwen3-235B-A22B sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 15.33–19.93 contaminationWarning Not reported rank 38 reportedConfidenceHalfWidth 2.3 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label Qwen3-235B-A22B Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/37` | | | | | | 12 | [gpt-oss-120b](https://benchize.com/models/gpt-oss-120b.md) | 15.17source score | CI 13.01–17.33 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 15.17 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gpt-oss-120b sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 13.01–17.33 contaminationWarning Potential contamination warning: This model was evaluated after the public release of MultiNRC, allowing model builder access to the prompts and solutions. rank 40 reportedConfidenceHalfWidth 2.16 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label gpt-oss-120b Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/39` | | | | | | 13 | [GPT-4o](https://benchize.com/models/gpt-4o.md) | 12.42source score | CI 10.43–14.41 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 12.42 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration GPT-4o sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 10.43–14.41 contaminationWarning Not reported rank 41 reportedConfidenceHalfWidth 1.99 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label GPT-4o Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/40` | | | | | | 14 | [gpt-oss-20b](https://benchize.com/models/gpt-oss-20b.md) | 10.43source score | CI 8.59–12.27 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 10.43 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gpt-oss-20b sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 8.59–12.27 contaminationWarning Potential contamination warning: This model was evaluated after the public release of MultiNRC, allowing model builder access to the prompts and solutions. rank 42 reportedConfidenceHalfWidth 1.84 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label gpt-oss-20b Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/41` | | | | | | 15 | [Llama 4 Maverick](https://benchize.com/models/llama-4-maverick.md) | 8.44source score | CI 6.76–10.12 | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 8.44 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Llama-4-Maverick sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 6.76–10.12 contaminationWarning Not reported rank 44 reportedConfidenceHalfWidth 1.68 ### Source [Original source](https://labs.scale.com/leaderboard/multinrc) Source scale Original model label Llama-4-Maverick Observed 2026-10-21T16:35:36.736Z Evidence SHA-256 `a4931a13575ec92cd4993e96d4e87f8ef58149c485b7b54bb8558d7762620f8a` Source locator `/rows/43` | | | | | --- Cite as: MultiNRC — Benchize, https://benchize.com/benchmarks/scale%3Amultinrc (retrieved 2026-10-21)