Vals Multimodal Index
A weighted composite of the multimodal tasks in the independent suite, intended as a single number for image-and-document work.
CompositeVals AI33 models run by Vals AIUpdated 11 August 2026
How it is run
Composites inherit every flaw of their components and hide which one moved.
What it cannot tell you
Use it to shortlist, never to decide. Read the component benchmarks before acting on a difference of a few points.
- Leader
- Claude Fable 574.15% · Vals AI's own list
- Top-ten spread
- 6.79%Leader minus tenth place, or last place in smaller fields.
- Median
- 60.57%of the 33 rows published
- No published result
- 29of 46 models on this site
Independent leaderboard · Vals AI
Top 33 of 33 runs by Vals AI, one harness.
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5 | 74.15% |
| 2 | Claude Opus 5 | 73.90% |
| 3 | Kimi K3 | 73.42% |
| 4 | GPT-5.6 Sol | 72.64% |
| 5 | Claude Opus 4.8 | 70.89% |
| 6 | Muse Spark 1.2 | 69.80% |
| 7 | GPT-5.6 Luna | 69.06% |
| 8 | Claude Sonnet 5 | 68.83% |
| 9 | GPT-5.5 | 68.07% |
| 10 | Claude Opus 4.7 | 67.36% |
| 11 | Muse Spark 1 1 | 66.74% |
| 12 | Qwen3.8 Max | 65.39% |
| 13 | Gemini 3.6 Flash | 65.08% |
| 14 | GPT-5.6 Terra | 65.07% |
| 15 | Grok 4.5 | 63.42% |
| 16 | Gemini 3.5 Flash | 62.93% |
| 17 | Claude Sonnet 4 6 | 60.57% |
| 18 | MiniMax M3 | 59.97% |
| 19 | Kimi K2.6 | 56.43% |
| 20 | Gemini 3.1 Pro | 56.07% |
| 21 | Gpt 5.4 Mini 2026 03 17 | 54.22% |
| 22 | Qwen3.7 Plus | 53.89% |
| 23 | Mimo V2.5 | 52.77% |
| 24 | Gemini 3 Flash Preview | 52.19% |
| 25 | Qwen3.6 Plus | 51.52% |
| 26 | Inkling Small | 50.05% |
| 27 | Inkling | 49.40% |
| 28 | Gpt 5.4 Nano 2026 03 17 | 47.64% |
| 29 | Grok 4.3 | 43.29% |
| 30 | Claude Haiku 4 5 20251001 Thinking | 42.88% |
| 31 | Gemini 3.1 Flash Lite Preview | 41.35% |
| 32 | Grok 4.20 0309 Reasoning | 39.06% |
| 33 | Mistral Medium 3.5 | 34.77% |
No published result tracked
29 of the 46 models on this site have no published number on this benchmark, so their cells in the home table stay blank rather than reading as zero.
Step 5 PreviewGPT-6 LunaGLM 5.3 FlashQwen3.8 2.4T A95BQwen3.8-Flash-NextMiMo-V2.6-FlashQwen3.8 Max (0902)Grok 4.6Gemini 3.7 FlashGPT-5.4DeepSeek V4.1 Flash 1017GLM-LegalClaude Haiku 5.5BeamGemini 4 ArgonClaude Sonnet 5.5Claude Opus 5.5Claude Fable 5.1GPT-6 AstraGPT-6 SolGPT-6.1 SolGemini 3.8 FlashGrok 4.7DeepSeek V4.1 FlashDeepSeek V4 ProGLM 5.3Muse Spark 1.3 MaxHunyuan Hy4 PreviewMiMo V2.6 Pro
Read this with
Benchmarks in the same category, which usually disagree with this one on purpose.
Sources
Every page used for this benchmark, read on 17 October 2026.
- Benchmark site or the page that publishes itwww.vals.ai/benchmarks/vals_multimodal_index
- Independent resultswww.vals.ai/benchmarks/vals_multimodal_index