Comparing 4
Browse all 30 models on the map or in the table.
GPT-6 LunavsClaude Haiku 5.5vsGLM 5.3 FlashvsGemini 3.8 Flash
Send one prompt to every model and read the answers side by side, with speed and cost for each.
Shareable link:
Before you run
Highlighted values are the best among your picks.
GPT-6 LunaOpen AINEW OpenAI's most efficient model for focused, high-volume tasks | Claude Haiku 5.5AnthropicNEW Fast Claude model for responsive assistance, classification, and lightweight agents | GLM 5.3 FlashZ.ai· open weights Native multimodal GLM model for efficient coding and long-horizon agent tasks | Gemini 3.8 FlashGoogle Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows | |
|---|---|---|---|---|
| Intelligence | ||||
| Intelligence index | 38.1 | 43.4 | 41.8 | 40.9 |
| Price | ||||
| Input, $ per 1M | $0.10 | $0.10 | $0.15 | $0.75 |
| Output, $ per 1M | $0.50 | $0.50 | $0.50 | $3.75 |
| Typical message (500 in / 250 out) | $0.00017 | $0.00017 | $0.00020 | $0.0013 |
| Repeated input, cached | $0.01 · 10× cheaper | $0.01 · 10× cheaper | $0.03 · 5× cheaper | $0.07 · 10× cheaper |
| Prompts over 200K tokens | $0.20 / $0.75 | Same price | Same price | Same price |
| Limits | ||||
| Context window | 1.05M | 1M | 1M | 1.05M |
| Longest answer | 128K tokens | 128K tokens | 131K tokens | 66K tokens |
| Knowledge cutoff | May 2026 | Jun 2026 | Not published | Not published |
| What it can take and do | ||||
| Accepts | TextImagePDF | TextImagePDF | TextImage | TextImagePDFAudioVideo |
| Reasoning effort | Off · Low · Med · High · X-high · Max | Low · Med · High · X-high · Max | Low · High · Max | Low · Med · High |
| Tool use | Yes | Yes | Yes | Yes |
| Guaranteed JSON output | Yes | Yes | Yes | Yes |
| Availability | ||||
| Released | 22 Sept 2026 | 7 Oct 2026 | 26 Aug 2026 | 2 Sept 2026 |
| Open weights | No | No | Yes, can self-host | No |
| API providers offering it | 36 | 33 | 76 | 29 |