# Best Value Multimodal AI Model in 2026 — Cost-Adjusted Rankings

> Top AI models ranked by multimodal benchmark performance per dollar. Cost-adjusted rankings using MMMU-Pro and OfficeQA Pro.

Multimodal workloads — processing images, charts, documents, and screenshots — often involve large inputs that drive up token costs quickly. This ranking divides each model's weighted multimodal score (MMMU-Pro, OfficeQA-Pro) by output token price. For document processing pipelines and visual AI applications running at scale, the value leaders here offer the best multimodal reasoning per dollar spent.

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

Last updated: October 8, 2026

## Rankings

| Rank | Model | Creator | Type | Context | Score/$ | Score | Output $/1M |
|------|-------|---------|------|---------|---------|-------|-------------|
| 1 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | OpenAI | Proprietary | 1.05M | 56.75 | 68.1 | $1.2 |
| 2 | [MiniMax M3](/models/minimax-m3) | MiniMax | Open Weight | 1M | 42.92 | 51.5 | $1.2 |
| 3 | [Inkling-Small](/models/inkling-small) | Thinking Machines Lab | Open Weight | 1M | 34.58 | 49.8 | $1.44 |
| 4 | [Grok 4.3](/models/grok-4-3) | xAI | Proprietary | 1M | 26.68 | 66.7 | $2.5 |
| 5 | [Step 5 Preview](/models/step-5-preview) | StepFun | Pending | 1M | 22.67 | 61.2 | $2.7 |
| 6 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | Google | Proprietary | 1M | 22.37 | 83.9 | $3.75 |
| 7 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | Google | Proprietary | 1M | 22.29 | 83.6 | $3.75 |
| 8 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | Open Weight | 256K | 22.27 | 66.8 | $3 |
| 9 | [GPT-5.4 nano](/models/gpt-5-4-nano) | OpenAI | Proprietary | 400K | 19.92 | 24.9 | $1.25 |
| 10 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | Open Weight | 128K | 17.75 | 63.9 | $3.6 |
| 11 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | Open Weight | 256K | 16.23 | 64.9 | $4 |
| 12 | [Grok 4.20](/models/grok-4-20-beta) | xAI | Proprietary | 2M | 14.24 | 35.6 | $2.5 |
| 13 | [GPT-5.4 mini](/models/gpt-5-4-mini) | OpenAI | Proprietary | 400K | 12.96 | 58.3 | $4.5 |
| 14 | [Inkling](/models/inkling) | Thinking Machines Lab | Open Weight | 1M | 10.6 | 49.6 | $4.68 |
| 15 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | Google | Proprietary | 1M | 9.76 | 87.8 | $9 |
| 16 | [Claude Sonnet 5](/models/claude-sonnet-5) | Anthropic | Proprietary | 1M | 7.84 | 78.4 | $10 |
| 17 | [Gemini 3.1 Pro](/models/gemini-3-1-pro) | Google | Proprietary | 1M | 6.68 | 80.1 | $12 |
| 18 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | OpenAI | Proprietary | 1.05M | 6.52 | 78.2 | $12 |
| 19 | [Gemini 3 Pro](/models/gemini-3-pro) | Google | Proprietary | 2M | 6.27 | 75.2 | $12 |
| 20 | [Kimi K3](/models/kimi-k3) | Moonshot AI | Pending | 1.05M | 5.96 | 89.4 | $15 |
| 21 | [GPT-5.2](/models/gpt-5-2) | OpenAI | Proprietary | 400K | 4.81 | 67.3 | $14 |
| 22 | [GPT-5.4](/models/gpt-5-4) | OpenAI | Proprietary | 1.05M | 4.62 | 69.3 | $15 |
| 23 | [Claude Opus 5.5](/models/claude-opus-5-5) | Anthropic | Proprietary | 1M | 4.44 | 88.8 | $20 |
| 24 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | OpenAI | Proprietary | 1.05M | 4.43 | 88.6 | $20 |
| 25 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | Anthropic | Proprietary | 200K | 3.67 | 55.1 | $15 |
| 26 | [Claude Opus 5](/models/claude-opus-5) | Anthropic | Proprietary | null | 3.55 | 88.8 | $25 |
| 27 | [Claude Opus 4.8](/models/claude-opus-4-8) | Anthropic | Proprietary | 1M | 3.51 | 87.7 | $25 |
| 28 | [Claude Opus 4.6](/models/claude-opus-4-6) | Anthropic | Proprietary | 1M | 2.42 | 60.6 | $25 |
| 29 | [GPT-5.5](/models/gpt-5-5) | OpenAI | Proprietary | 1M | 2.38 | 71.4 | $30 |
| 30 | [Claude Opus 4.7 (Adaptive)](/models/claude-opus-4-7-adaptive) | Anthropic | Proprietary | 1M | 2 | 50.1 | $25 |
| 31 | [Command A+](/models/command-a-plus) | Cohere | Open Weight | 128K | 1.61 | 16.1 | $10 |
| 32 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | Proprietary | 200K | 0.98 | 24.5 | $25 |
| — | [Mistral Large 4](/models/mistral-large-4) | Mistral | Pending | 1M | — | — | — |

## Key Takeaways

- Top model: [GPT-5.6 Luna](/models/gpt-5-6-luna) with a Score/$ of 56.75 (score 68.1, $1.2/1M output tokens)
- Best open-weight option: [MiniMax M3](/models/minimax-m3) at #2
- Models included: 33 (32 ranked, 1 listed without a rank)

## Compare the Leaders

- [GPT-5.6 Luna vs MiniMax M3](/compare/gpt-5-6-luna-vs-minimax-m3)