| Rank | Model | Provider | Score | Benchmarks | Inference | Agentic | Programming | Value | Price |
|---|---|---|---|---|---|---|---|---|---|
| 81 | Step3-VL-10B step3-vl-10b multimodalvisionmulti-input reasoning | StepFun | 47.4 overall | 47.4 | 0.0 | 0.0 | 0.0 | 0.0 | N/A |
| 82 | MiniMax M2.7 minimax-m2.7 codeprogrammingtool use | MiniMax | 47.3 overall | 0.0 | 52.8 | 50.8 | 35.9 | 55.0 | $0.3 in / $1.2 out |
| 83 | Mercury 2 mercury-2 codeprogrammingtool use | Inception | 47.2 overall | 44.6 | 72.5 | 0.0 | 22.3 | 69.2 | $0.25 in / $0.75 out |
| 84 | Llama 4 Maverick llama-4-maverick multimodalvisionmulti-input reasoning | Meta | 47.2 overall | 35.5 | 58.9 | 0.0 | 0.0 | 61.2 | $0.17 in / $0.6 out |
| 85 | Mistral Small 4 mistral-small-latest multimodalvisionmulti-input reasoning | Mistral AI | 47.0 overall | 34.8 | 55.9 | 0.0 | 0.0 | 66.9 | |
| 86 | GLM-4.7 glm-4.7 multimodalvisionmulti-input reasoning | Zhipu AI | 46.8 overall | 63.2 | 52.8 | 28.2 | 44.5 | 40.6 | $0.6 in / $2.2 out |
| 87 | Gemini 2.0 Flash-Lite gemini-2.0-flash-lite multimodalvisionmulti-input reasoning | Google | 46.8 overall | 25.7 | 63.2 | 0.0 | 0.0 | 79.7 | |
| 88 | Gemini 2.0 Flash Thinking gemini-2.0-flash-thinking multimodalvisionmulti-input reasoning | Google | 46.7 overall | 46.7 | 0.0 | 0.0 | 0.0 | 0.0 | |
| 89 | GLM-5 glm-5 codeprogrammingtool use | Zhipu AI | 46.2 overall | 0.0 | 22.1 | 51.3 | 65.3 | 30.2 | $1 in / $3.2 out |
| 90 | Devstral Small 1.1 devstral-small-2507 codeprogrammingtool use | Mistral AI | 46.1 overall | 0.0 | 64.5 | 0.0 | 15.0 | 85.0 | $0.1 in / $0.3 out |
| 91 | DeepSeek-V3.2 deepseek-v3.2 codeprogrammingtool use | DeepSeek | 45.8 overall | 58.1 | 52.5 | 16.6 | 45.9 | 70.0 | $0.26 in / $0.38 out |
| 92 | LongCat-Flash-Thinking-2601 longcat-flash-thinking-2601 codeprogrammingtool use | Meituan | 45.6 overall | 56.3 | 51.9 | 30.8 | 38.0 | 57.7 | |
| 93 | o4-mini o4-mini multimodalvisionmulti-input reasoning | OpenAI | 45.5 overall | 48.8 | 70.7 | 38.2 | 32.7 | 41.9 | $1.1 in / $4.4 out |
| 94 | Claude Sonnet 4 claude-sonnet-4-20250514 multimodalvisionmulti-input reasoning | Anthropic | 44.9 overall | 41.0 | 0.0 | 49.4 | 44.9 | 0.0 | |
| 95 | GPT-5.1 Codex gpt-5.1-codex multimodalvisionmulti-input reasoning | OpenAI | 44.9 overall | 0.0 | 48.6 | 0.0 | 51.2 | 25.1 | $1.25 in / $10 out |
| 96 | Gemini 2.5 Pro Preview 06-05 gemini-2.5-pro-preview-06-05 multimodalvisionmulti-input reasoning | Google | 44.7 overall | 51.7 | 63.2 | 0.0 | 30.0 | 27.9 | |
| 97 | Qwen3 VL 235B A22B Thinking qwen3-vl-235b-a22b-thinking multimodalvisionmulti-input reasoning | Alibaba Cloud / Qwen Team | 44.4 overall | 37.9 | 66.8 | 40.2 | 0.0 | 37.2 | |
| 98 | Grok Code Fast 1 grok-code-fast-1 codeprogrammingtool use | xAI | 44.2 overall | 0.0 | 47.2 | 0.0 | 39.7 | 49.5 | $0.2 in / $1.5 out |
| 99 | GPT-5.4 Mini gpt-5.4-mini texttext-to-textlanguage | OpenAI | 44.1 overall | 57.4 | 77.4 | 27.1 | 26.9 | 32.8 | |
| 100 | Qwen2.5-Coder 32B Instruct qwen-2.5-coder-32b-instruct textinference | Alibaba Cloud / Qwen Team | 44.1 overall | 0.0 | 20.9 | 0.0 | 0.0 | 81.2 | $0.09 in / $0.09 out |
81
Step3-VL-10B
StepFun
47.4
N/A
82
MiniMax M2.7
MiniMax
47.3
$0.3 in / $1.2 out
83
Mercury 2
Inception
47.2
$0.25 in / $0.75 out
84