# Vision-Language Reasoning on VL Reasoning Benchmarks Source: [Vision-Language Reasoning on VL Reasoning Benchmarks]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:51:21.107Z Evaluation of model performance on the VL Reasoning Benchmarks suite for Vision-Language Reasoning tasks, reporting scores across MVista, MVerse, and MVision components. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Vision-Language Reasoning | | Dataset (including any stated split) | VL Reasoning Benchmarks | | Metric | MVista Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 28 of 28 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [MVista Score](): Higher is better; unit: PERCENT. - [MVerse Score](): Higher is better; unit: PERCENT. - [MVision Score](): Higher is better; unit: PERCENT. - [Average Score](): Higher is better; unit: PERCENT. - [MMMU Score](): Higher is better; unit: PERCENT. - [R1-OV Score](): Higher is better; unit: PERCENT. - [MMStar Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | MVista Score | Result date | Report paper | | --- | --- | --- | --- | | FRISM | 74 | Not recorded | 697c3aefa8b01e76cc8da38e | | FRISM | 74 | Not recorded | 697c3aefa8b01e76cc8da38e | | IP-Merging | 73.4 | Not recorded | 697c3aefa8b01e76cc8da38e | | Base | 73.3 | Not recorded | 697c3aefa8b01e76cc8da38e | | IP-Merging | 73.2 | Not recorded | 697c3aefa8b01e76cc8da38e | | DARE | 73.1 | Not recorded | 697c3aefa8b01e76cc8da38e | | IP-Merging | 72.3 | Not recorded | 697c3aefa8b01e76cc8da38e | | IP-Merging | 72.2 | Not recorded | 697c3aefa8b01e76cc8da38e | | Base | 72 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 71.4 | Not recorded | 697c3aefa8b01e76cc8da38e | | FRISM | 69.9 | Not recorded | 697c3aefa8b01e76cc8da38e | | DARE | 69.6 | Not recorded | 697c3aefa8b01e76cc8da38e | | Ties-Merging | 69 | Not recorded | 697c3aefa8b01e76cc8da38e | | DARE | 68.9 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 68.2 | Not recorded | 697c3aefa8b01e76cc8da38e | | Base | 68.1 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 67.6 | Not recorded | 697c3aefa8b01e76cc8da38e | | Ties-Merging | 67.4 | Not recorded | 697c3aefa8b01e76cc8da38e | | IP-Merging | 67.4 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 67.2 | Not recorded | 697c3aefa8b01e76cc8da38e | | Ties-Merging | 66 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 63.5 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 34.7 | Not recorded | 697c3aefa8b01e76cc8da38e | | FRISM | 32.1 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 31.7 | Not recorded | 697c3aefa8b01e76cc8da38e | | TA | 31.4 | Not recorded | 697c3aefa8b01e76cc8da38e | | IP-Merging | 29.8 | Not recorded | 697c3aefa8b01e76cc8da38e | | Base | 29 | Not recorded | 697c3aefa8b01e76cc8da38e | ### FRISM MVista Score: 74. Result ID: 69977dfe5698665ec4d2861b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### FRISM MVista Score: 74. Result ID: 69977dfe5698665ec4d28629. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"InternVL3-8B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### IP-Merging MVista Score: 73.4. Result ID: 69977dfe5698665ec4d2861a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B","T":0.4}. This record does not assert independent reproduction. ### Base MVista Score: 73.3. Result ID: 69977dfe5698665ec4d28612. [Reporting paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### IP-Merging MVista Score: 73.2. Result ID: 69977dfe5698665ec4d28619. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B","T":0.3}. This record does not assert independent reproduction. ### DARE MVista Score: 73.1. Result ID: 69977dfe5698665ec4d28617. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### IP-Merging MVista Score: 72.3. Result ID: 69977dfe5698665ec4d28628. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"InternVL3-8B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### IP-Merging MVista Score: 72.2. Result ID: 69977dfe5698665ec4d28618. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B","T":0.2}. This record does not assert independent reproduction. ### Base MVista Score: 72. Result ID: 69977dfe5698665ec4d28624. [Reporting paper]() Recorded evaluation configuration: {"Merged Models":"InternVL3-8B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### TA MVista Score: 71.4. Result ID: 69977dfe5698665ec4d28613. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B","lambda":0.1}. This record does not assert independent reproduction. ### FRISM MVista Score: 69.9. Result ID: 69977dfe5698665ec4d28623. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B"}. This record does not assert independent reproduction. ### DARE MVista Score: 69.6. Result ID: 69977dfe5698665ec4d28627. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"InternVL3-8B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### Ties-Merging MVista Score: 69. Result ID: 69977dfe5698665ec4d28620. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B"}. This record does not assert independent reproduction. ### DARE MVista Score: 68.9. Result ID: 69977dfe5698665ec4d28621. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B"}. This record does not assert independent reproduction. ### TA MVista Score: 68.2. Result ID: 69977dfe5698665ec4d28625. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"InternVL3-8B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### Base MVista Score: 68.1. Result ID: 69977dfe5698665ec4d2861c. [Reporting paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B"}. This record does not assert independent reproduction. ### TA MVista Score: 67.6. Result ID: 69977dfe5698665ec4d2861d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B","lambda":0.1}. This record does not assert independent reproduction. ### Ties-Merging MVista Score: 67.4. Result ID: 69977dfe5698665ec4d28616. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### IP-Merging MVista Score: 67.4. Result ID: 69977dfe5698665ec4d28622. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B"}. This record does not assert independent reproduction. ### TA MVista Score: 67.2. Result ID: 69977dfe5698665ec4d2861e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B","lambda":0.15}. This record does not assert independent reproduction. ### Ties-Merging MVista Score: 66. Result ID: 69977dfe5698665ec4d28626. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"InternVL3-8B + DeepSeek-R1-distilled-Qwen-7B"}. This record does not assert independent reproduction. ### TA MVista Score: 63.5. Result ID: 69977dfe5698665ec4d2861f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"Qwen2.5-VL-7B-Instruct + OpenThinker3-7B","lambda":0.2}. This record does not assert independent reproduction. ### TA MVista Score: 34.7. Result ID: 69977dfe5698665ec4d28614. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B","lambda":0.15}. This record does not assert independent reproduction. ### FRISM MVista Score: 32.1. Result ID: 69977dfe5698665ec4d2860d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging Setup":"LLAVA-Next-8B with DeepThought-8B"}. This record does not assert independent reproduction. ### TA MVista Score: 31.7. Result ID: 69977dfe5698665ec4d28615. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merged Models":"ThinkLite-VL-7B + DeepSeek-R1-distilled-Qwen-7B","lambda":0.2}. This record does not assert independent reproduction. ### TA MVista Score: 31.4. Result ID: 69977dfe5698665ec4d2860b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging Setup":"LLAVA-Next-8B with DeepThought-8B","Method Type":"Task Arithmetic"}. This record does not assert independent reproduction. ### IP-Merging MVista Score: 29.8. Result ID: 69977dfe5698665ec4d2860c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging Setup":"LLAVA-Next-8B with DeepThought-8B"}. This record does not assert independent reproduction. ### Base MVista Score: 29. Result ID: 69977dfe5698665ec4d2860a. [Reporting paper]() Recorded evaluation configuration: {"Base Model":"LLAVA-Next-8B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.