Value ranking
Best Value LLMs for Multimodal in October 2026
As of October 2026, Gemma 4 26B A4B IT gives the most multimodal performance per dollar: 40.6 points at a blended $0.11 per million tokens. Gemma 4 31B IT is next.
Last verified
Category score per dollar of blended API price, among models in the top half of the multimodal ranking.
- Gemma 4 26B A4B IT 379.5 pts/$
- Gemma 4 31B IT 272.9 pts/$
- MiMo-V2.6-Flash 231.6 pts/$
- MiMo-V2.5 227.3 pts/$
- MiMo-V2-Omni 220.6 pts/$
- Claude Haiku 5.5 218.3 pts/$
- GPT-6 Luna 212.2 pts/$
- GLM-5.3-Flash 180.3 pts/$
- DeepSeek V4.1 Flash 148.8 pts/$
- GPT-5.6 Luna 94.9 pts/$
Full ranking
| # | Model | Score | Input $/M | Output $/M | Points per $ |
|---|---|---|---|---|---|
| 1 | Gemma 4 26B A4B IT | 40.6 | $0.0675 | $0.23 | 379.5 |
| 2 | Gemma 4 31B IT | 41.6 | $0.09 | $0.34 | 272.9 |
| 3 | MiMo-V2.6-Flash | 40.5 | $0.14 | $0.28 | 231.6 |
| 4 | MiMo-V2.5 | 39.8 | $0.14 | $0.28 | 227.3 |
| 5 | MiMo-V2-Omni | 38.6 | $0.14 | $0.28 | 220.6 |
| 6 | Claude Haiku 5.5 | 43.7 | $0.10 | $0.50 | 218.3 |
| 7 | GPT-6 Luna | 42.4 | $0.10 | $0.50 | 212.2 |
| 8 | GLM-5.3-Flash | 42.8 | $0.15 | $0.50 | 180.3 |
| 9 | DeepSeek V4.1 Flash | 39.1 | $0.15 | $0.60 | 148.8 |
| 10 | GPT-5.6 Luna | 42.7 | $0.20 | $1.20 | 94.9 |
| 11 | MiniMax-M3 | 40.2 | $0.30 | $1.20 | 76.5 |
| 12 | MiMo-V2.6-Pro | 40.8 | $0.43 | $0.87 | 75.1 |
| 13 | Gemini 3.1 Flash Lite | 39.4 | $0.25 | $1.50 | 70.0 |
| 14 | Inkling-Small | 39.1 | $0.45 | $1.20 | 61.3 |
| 15 | Qwen3.7 Plus | 41.8 | $0.40 | $1.60 | 59.7 |
| 16 | Gemini 2.5 Flash | 41.8 | $0.30 | $2.50 | 49.2 |
| 17 | Gemini 3.5 Flash Lite | 41.1 | $0.30 | $2.50 | 48.3 |
| 18 | Qwen3.5 27B | 39.4 | $0.30 | $2.40 | 47.7 |
| 19 | Kimi K2.5 | 41.1 | $0.45 | $2.25 | 45.7 |
| 20 | Gemini 3 Flash Preview | 45.5 | $0.50 | $3 | 40.4 |
| 21 | Qwen3.8 27B | 41.3 | $0.99 | $1.49 | 37.0 |
| 22 | Seed 2.0 Pro | 41.5 | $0.50 | $3 | 36.9 |
| 23 | Qwen3.5 122B-A10B | 39.6 | $0.40 | $3.20 | 36.0 |
| 24 | Qwen3-VL 235B-A22B | 39.8 | $0.70 | $2.80 | 32.5 |
| 25 | Qwen3.5 397B-A17B | 40.7 | $0.60 | $3.60 | 30.2 |
| 26 | Step 5 Preview | 41.0 | $1 | $2.70 | 28.8 |
| 27 | Gemini 3.8 Flash | 40.7 | $0.75 | $3.75 | 27.2 |
| 28 | Grok 4.20 Multi-Agent | 40.5 | $1.25 | $2.50 | 25.9 |
| 29 | Gemini 3.6 Flash | 38.5 | $0.75 | $3.75 | 25.7 |
| 30 | GPT-5.4 mini | 39.7 | $0.75 | $4.50 | 23.5 |
| 31 | Muse Spark 1.3 | 43.7 | $1.25 | $4.25 | 21.8 |
| 32 | Muse Spark 1.2 | 43.4 | $1.25 | $4.25 | 21.7 |
| 33 | GLM-5V-Turbo | 40.9 | $1.20 | $4 | 21.5 |
| 34 | Muse Spark 1.1 | 42.6 | $1.25 | $4.25 | 21.3 |
| 35 | o4-mini | 40.2 | $1.10 | $4.40 | 20.9 |
| 36 | Grok 4.6 | 43.6 | $2 | $6 | 14.5 |
| 37 | GPT-5 | 46.8 | $1.25 | $10 | 13.6 |
| 38 | Gemini 3.5 Flash | 45.7 | $1.50 | $9 | 13.5 |
| 39 | GPT-6.1 Sol | 52.7 | $2 | $10 | 13.2 |
| 40 | Gemini 2.5 Pro | 45.2 | $1.25 | $10 | 13.1 |
| 41 | GPT-5.1 | 44.8 | $1.25 | $10 | 13.0 |
| 42 | Claude Sonnet 5.5 | 51.5 | $2 | $10 | 12.9 |
| 43 | GPT-6 Sol | 47.6 | $2 | $10 | 11.9 |
| 44 | o3 | 41.4 | $2 | $8 | 11.8 |
| 45 | GPT-5.2 | 51.3 | $1.75 | $14 | 10.7 |
| 46 | Claude Sonnet 5 | 42.4 | $2 | $10 | 10.6 |
| 47 | GPT-5.6 Terra | 47.3 | $2 | $12 | 10.5 |
| 48 | GPT-5.4 | 43.7 | $2.50 | $15 | 7.8 |
| 49 | Claude Opus 5.5 | 57.8 | $4 | $20 | 7.2 |
| 50 | GPT-5.6 Sol | 48.6 | $4 | $20 | 6.1 |
| 51 | Claude Opus 5 | 50.8 | $5 | $25 | 5.1 |
| 52 | Claude Opus 4.8 | 42.9 | $5 | $25 | 4.3 |
| 53 | GPT-5.5 | 46.9 | $5 | $30 | 4.2 |
| 54 | Claude Opus 4.7 | 41.2 | $5 | $25 | 4.1 |
| 55 | GPT-6 Astra | 55.0 | $10 | $50 | 2.7 |
| 56 | Claude Fable 5.1 | 53.9 | $10 | $50 | 2.7 |
| 57 | Claude Fable 5 | 45.3 | $10 | $50 | 2.3 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Compare the leaders
Frequently asked questions
What is the most cost-effective AI model for multimodal?
As of October 2026, Gemma 4 26B A4B IT gives the most multimodal performance per dollar: 40.6 points at a blended $0.11 per million tokens. Gemma 4 31B IT is next.
What are the top 5 in this ranking?
Gemma 4 26B A4B IT, Gemma 4 31B IT, MiMo-V2.6-Flash, MiMo-V2.5, MiMo-V2-Omni, in that order, as of October 2026.
How is this list ranked?
By score per dollar: the model’s score divided by its blended API price (three parts input to one part output), among models in the top half by score.