# Best Multimodal LLMs in 2026

> AI models ranked for multimodal understanding — images, documents, charts, and grounded visual reasoning.

This page ranks models by BenchLM's multimodal & grounded category: image understanding, document and chart reading, and visually grounded reasoning. Note the scope: these scores measure how well a model understands images — not how well it generates them. Image-generation systems are a different product class and are not ranked here.

Canonical page: https://benchlm.ai/best/multimodal

Last updated: October 10, 2026

## Rankings

| Rank | Model | Creator | Type | Context | Score |
|------|-------|---------|------|---------|-------|
| 1 | [Kimi K3](/models/kimi-k3) | Moonshot AI | Pending | 1.05M | 92.6 |
| 2 | [Claude Opus 5](/models/claude-opus-5) | Anthropic | Proprietary | null | 92.2 |
| 3 | [Claude Opus 5.5](/models/claude-opus-5-5) | Anthropic | Proprietary | 1M | 92.2 |
| 4 | [Qwen3.8 Max](/models/qwen3-8-max) | Alibaba | Open Weight | 1M | 91.2 |
| 5 | [Claude Opus 4.8](/models/claude-opus-4-8) | Anthropic | Proprietary | 1M | 91.2 |
| 6 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | OpenAI | Proprietary | 1.05M | 89.8 |
| 7 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | Google | Proprietary | 1M | 87.8 |
| 8 | [Qwen3.8-Flash-Next](/models/qwen3-8-flash-next) | Alibaba | Open Weight | 262K | 87.7 |
| 9 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | Google | Proprietary | 1M | 87.3 |
| 10 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | Google | Proprietary | 1M | 87.2 |
| 11 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | OpenAI | Proprietary | 1.05M | 84.5 |
| 12 | [Muse Spark](/models/muse-spark) | Meta | Proprietary | 262K | 84.4 |
| 13 | [Qwen3.8-27B](/models/qwen3-8-27b) | Alibaba | Open Weight | 262K | 84.4 |
| 14 | [GLM-5.3-Flash](/models/glm-5-3-flash) | Z.AI | Open Weight | 1M | 84 |
| 15 | [Claude Sonnet 5](/models/claude-sonnet-5) | Anthropic | Proprietary | 1M | 82 |
| 16 | [Qwen3.7 Plus](/models/qwen3-7-plus) | Alibaba | Proprietary | 1M | 82 |
| 17 | [Gemini 3.1 Pro](/models/gemini-3-1-pro) | Google | Proprietary | 1M | 80.1 |
| 18 | [Gemini 3 Pro](/models/gemini-3-pro) | Google | Proprietary | 2M | 80 |
| 19 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | OpenAI | Proprietary | 1.05M | 79.6 |
| 20 | [Grok 4.3](/models/grok-4-3) | xAI | Proprietary | 1M | 78.8 |
| 21 | [Qwen3.6 Plus](/models/qwen3-6-plus) | Alibaba | Proprietary | 1M | 76.1 |
| 22 | [GPT-5.2](/models/gpt-5-2) | OpenAI | Proprietary | 400K | 75.9 |
| 23 | [GPT-5.5](/models/gpt-5-5) | OpenAI | Proprietary | 1M | 75.5 |
| 24 | [Step 5 Preview](/models/step-5-preview) | StepFun | Pending | 1M | 74.6 |
| 25 | [Claude Opus 4.6](/models/claude-opus-4-6) | Anthropic | Proprietary | 1M | 74.5 |
| 26 | [Gemma 4 31B](/models/gemma-4-31b) | Google | Open Weight | 256K | 74.3 |
| 27 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | Open Weight | 256K | 73.8 |
| 28 | [GPT-5.4 mini](/models/gpt-5-4-mini) | OpenAI | Proprietary | 400K | 73.8 |
| 29 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | Open Weight | 128K | 73.6 |
| 30 | [GPT-5.4](/models/gpt-5-4) | OpenAI | Proprietary | 1.05M | 73.1 |
| 31 | [MiMo-V2.5](/models/mimo-v2-5) | Xiaomi | Proprietary | 1M | 72.2 |
| 32 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | Open Weight | 256K | 70.4 |
| 33 | [Kimi K2.5 (Reasoning)](/models/kimi-k2-5-reasoning) | Moonshot AI | Proprietary | 128K | 69 |
| 34 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | Alibaba | Open Weight | 262K | 67.8 |
| 35 | [Inkling-Small](/models/inkling-small) | Thinking Machines Lab | Open Weight | 1M | 67.5 |
| 36 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | Open Weight | 262K | 67.5 |
| 37 | [Inkling](/models/inkling) | Thinking Machines Lab | Open Weight | 1M | 67.3 |
| 38 | [Qwen3.6-27B](/models/qwen3-6-27b) | Alibaba | Open Weight | 262K | 67.3 |
| 39 | [Gemma 4 26B A4B](/models/gemma-4-26b-a4b) | Google | Open Weight | 256K | 66.9 |
| 40 | [Qwen3.6-35B-A3B](/models/qwen3-6-35b-a3b) | Alibaba | Open Weight | 262K | 66.4 |
| 41 | [Muse Glimmer 30B](/models/muse-glimmer-30b) | Meta | Open Weight | 131K | 65.1 |
| 42 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | Open Weight | 262K | 64.6 |
| 43 | [MiniMax M3](/models/minimax-m3) | MiniMax | Open Weight | 1M | 60 |
| 44 | [Gemma 4 12B](/models/gemma-4-12b) | Google | Open Weight | 256K | 59.4 |
| 45 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | Anthropic | Proprietary | 200K | 58.7 |
| 46 | [Claude Opus 4.7 (Adaptive)](/models/claude-opus-4-7-adaptive) | Anthropic | Proprietary | 1M | 53.7 |
| 47 | [Grok 4.20](/models/grok-4-20-beta) | xAI | Proprietary | 2M | 53.4 |
| 48 | [GPT-5.4 nano](/models/gpt-5-4-nano) | OpenAI | Proprietary | 400K | 52.1 |
| 49 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | Proprietary | 200K | 48.6 |
| 50 | [Nemotron 3 Nano Omni 30B A3B](/models/nemotron-3-nano-omni-30b-a3b) | NVIDIA | Open Weight | 256K | 43.8 |
| 51 | [Command A+](/models/command-a-plus) | Cohere | Open Weight | 128K | 34.6 |
| 52 | [Gemma 4 E4B](/models/gemma-4-e4b) | Google | Open Weight | 128K | 22.2 |
| 53 | [Gemma 4 E2B](/models/gemma-4-e2b) | Google | Open Weight | 128K | 13.4 |
| — | [Mistral Large 4](/models/mistral-large-4) | Mistral | Pending | 1M | — |

## Key Takeaways

- Top model: [Kimi K3](/models/kimi-k3) with a score of 92.6
- Best open-weight option: [Qwen3.8 Max](/models/qwen3-8-max) at #4
- Models included: 54 (53 ranked, 1 listed without a rank)

## Compare the Leaders

- [Kimi K3 vs Claude Opus 5](/compare/claude-opus-5-vs-kimi-k3)