| Rank | Model | Provider | Score | Benchmarks | Inference | Agentic | Programming | Value | Price |
|---|---|---|---|---|---|---|---|---|---|
| 21 | Gemini 3.1 Pro gemini-3.1-pro-preview multimodalvisionmulti-input reasoning | Google | 52.6 Agentic | 72.3 | 58.3 | 52.6 | 60.9 | 22.7 | $2.5 in / $15 out |
| 22 | DeepSeek-V4-Pro-Max deepseek-v4-pro-max codeprogrammingtool use | DeepSeek | 51.0 Agentic | 64.9 | 84.8 | 51.0 | 52.0 | 45.1 | |
| 23 | Qwen3 VL 235B A22B Instruct qwen3-vl-235b-a22b-instruct multimodalvisionmulti-input reasoning | Alibaba Cloud / Qwen Team | 51.0 Agentic | 34.3 | 0.0 | 51.0 | 0.0 | 0.0 | |
| 24 | Claude Haiku 4.5 claude-haiku-4-5-20251001 multimodalvisionmulti-input reasoning | Anthropic | 50.8 Agentic | 30.8 | 53.3 | 50.8 | 53.9 | 45.6 | |
| 25 | Kimi K2-Thinking-0905 kimi-k2-thinking-0905 codeprogrammingtool use | Moonshot AI | 50.8 Agentic | 66.2 | 0.0 | 50.8 | 58.0 | 0.0 | |
| 26 | Qwen3-Coder 480B A35B Instruct qwen3-coder-480b-a35b-instruct codeprogrammingtool use | Alibaba Cloud / Qwen Team | 50.7 Agentic | 0.0 | 0.0 | 50.7 | 32.1 | 0.0 | |
| 27 | Claude Opus 4.6 claude-opus-4-6 multimodalvisionmulti-input reasoning | Anthropic | 49.8 Agentic | 74.6 | 28.3 | 49.8 | 71.5 | 8.0 | |
| 28 | Qwen3.7 Max qwen3.7-max multimodalvisionmulti-input reasoning | Alibaba Cloud / Qwen Team | 49.6 Agentic | 64.7 | 62.8 | 49.6 | 74.2 | 43.9 | $1.25 in / $3.75 out |
| 29 | Claude Sonnet 4 claude-sonnet-4-20250514 multimodalvisionmulti-input reasoning | Anthropic | 49.4 Agentic | 39.3 | 0.0 | 49.4 | 42.8 | 0.0 | |
| 30 | Claude 3.7 Sonnet claude-3-7-sonnet-20250219 multimodalvisionmulti-input reasoning | Anthropic | 49.1 Agentic | 42.3 | 0.0 | 49.1 | 37.7 | 0.0 | |
| 31 | GLM-5V-Turbo glm-5v-turbo multimodalvisionmulti-input reasoning | Zhipu AI | 49.1 Agentic | 0.0 | 0.0 | 49.1 | 0.0 | 0.0 | N/A |
| 32 | GPT-5.4 gpt-5.4 texttext-to-textlanguage | OpenAI | 48.9 Agentic | 70.5 | 37.2 | 48.9 | 50.3 | 18.5 | |
| 33 | LongCat-Flash-Chat longcat-flash-chat codeprogrammingtool use | Meituan | 48.1 Agentic | 26.0 | 0.0 | 48.1 | 36.6 | 0.0 | N/A |
| 34 | Qwen3.7-Plus qwen3.7-plus multimodalvisionmulti-input reasoning | Alibaba Cloud / Qwen Team | 48.1 Agentic | 62.9 | 62.8 | 48.1 | 61.1 | 69.5 | $0.32 in / $1.28 out |
| 35 | Kimi K2.6 kimi-k2.6 texttext-to-textlanguage | Moonshot AI | 47.2 Agentic | 63.5 | 32.9 | 47.2 | 69.4 | 46.3 | |
| 36 | Kimi K2.7 Code kimi-k2.7-code multimodalvisionmulti-input reasoning | Moonshot AI | 47.2 Agentic | 0.0 | 32.9 | 47.2 | 0.0 | 47.6 | |
| 37 | GPT-5.2 Pro gpt-5.2-pro-2025-12-11 multimodalvisionmulti-input reasoning | OpenAI | 46.1 Agentic | 64.4 | 0.0 | 46.1 | 0.0 | 0.0 | |
| 38 | MiniMax M2.1 minimax-m2.1 codeprogrammingtool use | MiniMax | 45.7 Agentic | 39.1 | 68.9 | 45.7 | 47.6 | 72.9 | $0.3 in / $1.2 out |
| 39 | Seed 2.0 Pro seed-2.0-pro multimodalvisionmulti-input reasoning | ByteDance | 44.8 Agentic | 66.8 | 23.2 | 44.8 | 56.0 | 54.9 | $0.5 in / $3 out |
| 40 | o1 o1-2024-12-17 multimodalvisionmulti-input reasoning | OpenAI | 44.7 Agentic | 41.9 | 0.0 | 44.7 | 5.6 | 0.0 | N/A |
21Google
Gemini 3.1 Pro
52.6
$2.5 in / $15 out
22
DeepSeek-V4-Pro-Max
DeepSeek
51.0
$1.6 in / $3.2 out
23
Qwen3 VL 235B A22B Instruct
Alibaba Cloud / Qwen Team
51.0
N/A
24