Ratings
Ratings
Every tool and model, highest score first. Open a name for the reason behind each number.
Products
38 products.
Product ratings
| Product | Capability | Reliability | Workflow fit | Pricing and value | Privacy and security | Support | Overall |
|---|---|---|---|---|---|---|---|
| GitHub Copilot | 7 | 7 | 8 | 6 | 5.5 | 6 | |
| AssemblyAI | 7 | 5.5 | 7.5 | 7.5 | 6 | 5.5 | |
| Deepgram | 7 | 5.5 | 7.5 | 7.5 | 5 | 5.5 | |
| Claude Code | 8 | 5 | 7 | 6 | 6.5 | 5.5 | |
| NotebookLM | 6 | 5 | 8 | 8 | 5.5 | 6 | |
| ElevenLabs | 8 | 6 | 7 | 6 | 3.5 | 5 | |
| Adobe Firefly | 7 | 5 | 8 | 6 | 6 | 5 | |
| OpenAI Codex | 6 | 6 | 7 | 8 | 6 | 5 | |
| ChatGPT | 7 | 5 | 8 | 6 | 5.5 | 5 | |
| Cursor | 7 | 6 | 8 | 5 | 6 | 5 | |
| Stable Diffusion | 7 | 5 | 7.5 | 6.5 | 5.5 | 5 | |
| Gemini | 7 | 5 | 7.5 | 6.5 | 5 | 5 | |
| Elicit | 7 | 5 | 7 | 6 | 6 | 5.5 | |
| Google Agent Development Kit | 7 | 6 | 5 | 8 | 5 | 6 | |
| Grammarly | 6.5 | 5 | 7.5 | 6.5 | 6 | 5 | |
| Microsoft Agent Framework | 6.5 | 5.5 | 6 | 8 | 5 | 6 | |
| Otter.ai | 6.5 | 5 | 7.5 | 6.5 | 5.5 | 4.5 | |
| Microsoft Copilot | 6.5 | 5 | 8 | 5.5 | 6 | 5 | |
| Windsurf | 6.5 | 5 | 7.5 | 6 | 5 | 5 | |
| LangGraph | 7 | 6 | 4 | 8 | 6.5 | 6 | |
| Midjourney | 8 | 5 | 6 | 6 | 3 | 4.5 | |
| Replit | 7 | 5 | 7.5 | 5 | 5 | 5 | |
| Claude | 6 | 4 | 8 | 6 | 6 | 5 | |
| Murf AI | 7 | 5 | 7 | 5 | 6 | 5 | |
| Notion AI | 6 | 5 | 7 | 6 | 7.5 | 5 | |
| Tableau | 6.5 | 5 | 7 | 5.5 | 5.5 | 6 | |
| Descript | 6 | 5 | 7 | 6 | 4 | 5 | |
| Grok | 6.5 | 4.5 | 7 | 5.5 | 4.5 | 4 | |
| Amazon Bedrock AgentCore | 7 | 6 | 4 | 6 | 7 | 5 | |
| Perplexity | 5 | 4 | 8 | 6 | 6 | 4.5 | |
| Zapier Agents | 5.5 | 4 | 8 | 5 | 5 | 5 | |
| Consensus | 4 | 3 | 8 | 7 | 5.5 | 4.5 | |
| Runway | 6 | 5 | 6 | 5 | 3 | 4 | |
| Databricks Genie | 6 | 5 | 5 | 5 | 4.5 | 5 | |
| Snowflake Cortex Analyst | 5 | 5 | 5 | 5 | 5.5 | 4.5 | |
| Copilot in Power BI | 5 | 5 | 5 | 4 | 4 | 6 | |
| Salesforce Agentforce | 5.5 | 4 | 5 | 3 | 7 | 5 | |
| Gemini Code Assist | 5 | 4 | 6 | 2 | 7 | 3 |
Models
15 models.
Model ratings
| Model | Output quality | Reliability | Speed | Ease of use | Value | Overall |
|---|---|---|---|---|---|---|
| Claude Sonnet 5.5 | 7 | 6 | 7 | 7 | 8 | |
| GPT-6 Luna | 5 | 6 | 8 | 7 | 9 | |
| Claude Opus 5.5 | 8 | 6 | 6 | 7 | 7 | |
| GPT-6.1 Sol | 7 | 6 | 6 | 7 | 8 | |
| Grok 4.7 | 7 | 6 | 6 | 7 | 8 | |
| Muse Spark 1.3 | 6 | 6 | 7 | 7 | 8 | |
| DeepSeek V4.1 Flash | 6 | 5 | 7 | 6 | 9 | |
| Gemini 3.8 Flash | 6 | 6 | 7 | 7 | 7 | |
| GLM-5.3 | 7 | 5 | 6 | 6 | 7 | |
| GPT-6 Astra | 8 | 6 | 5 | 7 | 5 | |
| Mistral Medium 3.5 | 6 | 6 | 6 | 6 | 7 | |
| Qwen3.8 Max | 6 | 6 | 6 | 6 | 7 | |
| Kimi K3.1 Pro | 7 | 5 | 5 | 5 | 8 | |
| Claude Fable 5.1 | 8 | 6 | 4 | 7 | 4 | |
| Kimi K3 | 7 | 5 | 5 | 5 | 5 |
Frequently asked questions
Where is the reason for a score?
Open the product or model. The rating section gives one reason for each dimension, next to the score.
What if there is little evidence for a dimension?
It is still rated, and the reason says so: for example, rated from documented features and stated limits because no independent test exists. A rating on thin evidence is never presented as a measurement.
How are model ratings worked out?
On output quality, reliability, speed, ease of use and value, from vendor pages and vendor-reported benchmarks that were not reproduced. Each model page gives the reason, the list price and the context window.