# AIモデル比較・検討
**URL:** https://forum.ficusonline.com/t/topic/553
**Category:** AI (LLM VLM Multimodal)
**Tags:** ai, llm
**Created:** [2026 年 4 月 9 日午前 2:32 UTC](https://forum.ficusonline.com/t/topic/553 "2026-04-09T02:32:03Z")
**Posts on this page:** 2
**Page:** 1
### Author:  [@tk-fuse](https://forum.ficusonline.com/u/tk-fuse)
#### Post date: [2026 年 4 月 9 日午前 2:32 UTC](https://forum.ficusonline.com/t/topic/553/1 "2026-04-09T02:32:03Z")
### AIモデル一覧
> **[Models – Hugging Face](https://huggingface.co/models)**
>
> Explore machine learning models.
> **[ModelScope 魔搭社区](https://modelscope.cn/home)**
>
> ModelScope——汇聚各领域先进的机器学习模型，提供模型探索体验、推理、训练、部署和应用的一站式服务。在这里，共建模型开源社区，发现、学习、定制和分享心仪的模型。
### AIモデル実行アプリケーション
> **[Model Catalog - LM Studio](https://lmstudio.ai/models)**
>
> The latest and greatest LLMs you can run on your computer.
> **[Ollama](https://ollama.com/search)**
>
> Search for models on Ollama.
> **[GitHub - PrismML-Eng/llama.cpp: LLM inference in C/C++](https://github.com/PrismML-Eng/llama.cpp)**
>
> LLM inference in C/C++
### AIモデル比較・動向
> **[LMSYS Org](https://www.lmsys.org/)**
>
> Large Model Systems Organization - developing large models and systems that are open, accessible, and scalable.
> **[Arena | Benchmark & Compare the Best AI Models](https://arena.ai/)**
>
> Chat with multiple AI models side-by-side. Compare ChatGPT, Claude, Gemini, and other top LLMs. Crowdsourced benchmarks and leaderboards.
---
### Author:  [@tk-fuse](https://forum.ficusonline.com/u/tk-fuse)
#### Post date: [2026 年 4 月 9 日午前 5:50 UTC](https://forum.ficusonline.com/t/topic/553/2 "2026-04-09T05:50:03Z")
## AIモデルランキング
* * *
チャット、コードベースの分野では、ChatGPTやGeminiを抑えClaudeのOpus 4.6の評価が一番高くなっています（あくまで投稿時点での評価です）。
* * *
### コーディング（プロプライエタリ + オープンソース）
View overall rankings across AI models on agentic coding tasks involving multi-step reasoning and tool use.
> **[Code AI Leaderboard - Best AI Models for Coding](https://arena.ai/leaderboard/code/webdev)**
>
> Chat with multiple AI models side-by-side. Compare ChatGPT, Claude, Gemini, and other top LLMs. Crowdsourced benchmarks and leaderboards.
| Rank | Model | Score | Price $/M | Context |
| --- | --- | --- | --- | --- |
| 1 | [claude-opus-4-6-thinking](https://www.anthropic.com/news/claude-opus-4-6) Anthropic · Proprietary | 1546+12/-12 | $5 / $25 | 1M |
| 2 | [claude-opus-4-6](https://www.anthropic.com/news/claude-opus-4-6) Anthropic · Proprietary | 1543+11/-11 | $5 / $25 | 1M |
| 3 | [claude-sonnet-4-6](https://www.anthropic.com/news/claude-sonnet-4-6) Anthropic · Proprietary | 1521+9/-9 | $3 / $15 | 1M |
| 4 | [claude-opus-4-5-20251101-thinking-32k](https://www.anthropic.com/news/claude-opus-4-5) Anthropic · Proprietary | 1491+7/-7 | $5 / $25 | 200K |
| 5 | [claude-opus-4-5-20251101](https://www.anthropic.com/news/claude-opus-4-5) Anthropic · Proprietary | 1465+7/-7 | $5 / $25 | 200K |
| 6 | [gpt-5.4-high (codex-harness)](https://platform.openai.com/docs/models/gpt-5.4) OpenAI · Proprietary | 1457+17/-17 | N/A | N/A |
| 7 | [gemini-3.1-pro-preview](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/) Google · Proprietary | 1456+9/-9 | $2 / $12 | 1M |
| 8 | [qwen3.6-plus-preview](https://chat.qwen.ai/?models=Qwen3.6-Plus-Preview) Alibaba · Proprietary | 1454+19/-19 Preliminary | $0.33 / $1.95 | 1M |
| 9 | [glm-5](https://huggingface.co/zai-org/GLM-5) Z.ai · MIT | 1441+10/-10 | $1 / $3.20 | 202.8K |
| 10 | [glm-4.7](https://huggingface.co/zai-org/GLM-4.7) Z.ai · MIT | 1439+10/-10 | $0.39 / $1.75 | 202.8K |
### コーディング（オープンソースに限定）
| Rank | Overall Rank | Model | Score | Price $/M | Context |
| --- | --- | --- | --- | --- | --- |
| 1 | 9 | [glm-5](https://huggingface.co/zai-org/GLM-5) Z.ai · MIT | 1441+10/-10 | $1 / $3.20 | 202.8K |
| 2 | 10 | [glm-4.7](https://huggingface.co/zai-org/GLM-4.7) Z.ai · MIT | 1439+10/-10 | $0.39 / $1.75 | 202.8K |
| 3 | 14 | [kimi-k2.5-thinking](https://www.kimi.com/blog/kimi-k2-5.html) Moonshot · Modified MIT | 1429+8/-8 | $0.60 / $3 | N/A |
| 4 | 17 | [kimi-k2.5-instant](https://www.kimi.com/blog/kimi-k2-5.html) Moonshot · Modified MIT | 1408+11/-11 | $0.38 / $1.72 | 262.1K |
| 5 | 20 | [minimax-m2.5](https://www.minimax.io/news/minimax-m25) MiniMax · Modified MIT | 1396+8/-8 | $0.12 / $0.99 | 196.6K |
| 6 | 22 | [minimax-m2.1-preview](https://www.minimax.io/news/minimax-m21) MiniMax · MIT | 1391+8/-8 | $0.27 / $0.95 | 196.6K |
| 7 | 26 | [qwen3.5-397b-a17b](https://huggingface.co/Qwen/Qwen3.5-397B-A17B) Alibaba · Apache 2.0 | 1386+9/-9 | $0.39 / $2.34 | 262.1K |
| 8 | 31 | [deepseek-v3.2-thinking](https://api-docs.deepseek.com/news/news250929) DeepSeek · MIT | 1368+8/-8 | $0.26 / $0.38 | 163.8K |
| 9 | 32 | [qwen3.5-122b-a10b](https://huggingface.co/Qwen/Qwen3.5-122B-A10B) Alibaba · Apache 2.0 | 1362+10/-10 | $0.26 / $2.08 | 262.1K |
| 10 | 33 | [glm-4.6](https://docs.z.ai/guides/llm/glm-4.6) Z.ai · MIT | 1354+9/-9 | $0.39 / $1.90 | 204.8K |
* * *
### チャット（テキストベース、プロプライエタリ + オープンソース）
> **[LLM Leaderboard - Best Text & Chat AI Models Compared](https://arena.ai/leaderboard/text)**
>
> View overall rankings across various AI models in text-to-text tasks across math, coding, creative writing, and other open-ended domains.
| Rank | Model | Score | Price $/M | Context |
| --- | --- | --- | --- | --- |
| 1 | [claude-opus-4-6-thinking](https://www.anthropic.com/news/claude-opus-4-6) Anthropic · Proprietary | 1503±5 | $5 / $25 | 1M |
| 2 | [claude-opus-4-6](https://www.anthropic.com/news/claude-opus-4-6) Anthropic · Proprietary | 1497±5 | $5 / $25 | 1M |
| 3 | [gemini-3.1-pro-preview](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/) Google · Proprietary | 1493±5 | $2 / $12 | 1M |
| 4 | [grok-4.20-beta1](https://grok.com/) xAI · Proprietary | 1490±7 | N/A | N/A |
| 5 | [gemini-3-pro](http://aistudio.google.com/app/prompts/new_chat?model=gemini-3-pro-preview) Google · Proprietary | 1486±4 | $2 / $12 | 1M |
| 6 | [gpt-5.4-high](https://platform.openai.com/docs/models/gpt-5.4) OpenAI · Proprietary | 1484±7 | $2.50 / $15 | 1.1M |
| 7 | [grok-4.20-beta-0309-reasoning](https://docs.x.ai/developers/models/grok-4.20-beta-0309-reasoning) xAI · Proprietary | 1480±6 | $2 / $6 | 2M |
| 8 | [gpt-5.2-chat-latest-20260210](https://developers.openai.com/api/docs/models/gpt-5.2-chat-latest) OpenAI · Proprietary | 1477±5 | $1.75 / $14 | 128K |
| 9 | [grok-4.20-multi-agent-beta-0309](https://docs.x.ai/developers/models/grok-4.20-multi-agent-beta-0309) xAI · Proprietary | 1475±6 | $2 / $6 | 2M |
| 10 | [gemini-3-flash](https://blog.google/products/gemini/gemini-3-flash) Google · Proprietary | 1474±4 | $0.50 / $3 | 1M |
### チャット（テキストベース、オープンソース限定）
| Rank | Overall Rank | Model | Score | Price $/M | Context |
| --- | --- | --- | --- | --- | --- |
| 1 | 14 | [glm-5.1](https://huggingface.co/zai-org/GLM-5.1) Z.ai · MIT | 1467±9 | $1.40 / $4.40 | 202.8K |
| 2 | 23 | [glm-5](https://huggingface.co/zai-org/GLM-5) Z.ai · MIT | 1456±6 | $1 / $3.20 | 202.8K |
| 3 | 25 | [kimi-k2.5-thinking](https://www.kimi.com/blog/kimi-k2-5.html) Moonshot · Modified MIT | 1452±5 | $0.60 / $3 | N/A |
| 4 | 29 | [gemma-4-31b](https://aistudio.google.com/app/prompts/new_chat?model=gemma-4-31b-it) Google · Apache 2.0 | 1451±8 Preliminary | $0.14 / $0.40 | 262.1K |
| 5 | 33 | [qwen3.5-397b-a17b](https://huggingface.co/Qwen/Qwen3.5-397B-A17B) Alibaba · Apache 2.0 | 1448±5 | $0.39 / $2.34 | 262.1K |
| 6 | 39 | [glm-4.7](https://huggingface.co/zai-org/GLM-4.7) Z.ai · MIT | 1443±6 | $0.39 / $1.75 | 202.8K |
| 7 | 43 | [gemma-4-26b-a4b](https://aistudio.google.com/app/prompts/new_chat?model=gemma-4-26b-a4b-it) Google · Apache 2.0 | 1438±8 Preliminary | N/A | N/A |
| 8 | 47 | [kimi-k2.5-instant](https://www.kimi.com/blog/kimi-k2-5.html) Moonshot · Modified MIT | 1433±7 | $0.38 / $1.72 | 262.1K |
| 9 | 50 | [kimi-k2-thinking-turbo](https://huggingface.co/moonshotai/Kimi-K2-Thinking) Moonshot · Modified MIT | 1430±4 | $1.15 / $8 | 262.1K |
| 10 | 53 | [glm-4.6](https://docs.z.ai/guides/llm/glm-4.6) Z.ai · MIT | 1426±4 | $0.39 / $1.90 | 204.8K |