THE THEUS UNIVERSE
One account. Every intelligence.
Find the model for your next idea. Compare capabilities, usage and plans in one place.
All 199. Search, filter, compare.
- Anthropic
Claude Opus 4.8
A code model in the Theus catalogue.
- Output usage
- ×5TT / token
- Context
- 200K
- Included from
- Pro
- Anthropic
Claude Opus 5
A multimodal model in the Theus catalogue.
- Output usage
- ×5TT / token
- Context
- 1M
- Included from
- Max 10x
- Anthropic
Claude Sonnet 5
A code model in the Theus catalogue.
- Output usage
- ×3TT / token
- Context
- 200K
- Included from
- Pro
- Google
Gemini 3.1 Pro
A multimodal model in the Theus catalogue.
- Output usage
- ×3.6TT / token
- Context
- 1M
- Included from
- Pro
- OpenAI
GPT-5.5
A reasoning model in the Theus catalogue.
- Output usage
- ×9TT / token
- Context
- 1.1M
- Included from
- Pro
- OpenAI
GPT-5.6 Sol
A reasoning model in the Theus catalogue.
- Output usage
- ×9TT / token
- Context
- 400K
- Included from
- Pro
- Anthropic
Claude Opus 4.5
A multimodal model in the Theus catalogue.
- Output usage
- ×5TT / token
- Context
- 200K
- Included from
- Max 10x
- Anthropic
Claude Opus 4.6
A multimodal model in the Theus catalogue.
- Output usage
- ×5TT / token
- Context
- 1M
- Included from
- Max 10x
- Anthropic
Claude Opus 4.7
A code model in the Theus catalogue.
- Output usage
- ×5TT / token
- Context
- 1M
- Included from
- Pro
- Anthropic
Claude Sonnet 4.6
A multimodal model in the Theus catalogue.
- Output usage
- ×3TT / token
- Context
- 1M
- Included from
- Max 10x
- DeepSeek
DeepSeek-V4 Pro
A code model in the Theus catalogue.
- Output usage
- ×1TT / token
- Context
- 1M
- Included from
- Pro
- Google
Gemini 3.5 Flash
A multimodal model in the Theus catalogue.
- Output usage
- ×1.8TT / token
- Context
- 1M
- Included from
- Max
- Google
Gemini 3.6 Flash
A multimodal model in the Theus catalogue.
- Output usage
- ×1.5TT / token
- Context
- 1M
- Included from
- Max
- Z.ai
GLM 5.3
A code model in the Theus catalogue.
- Output usage
- ×1TT / token
- Context
- 1M
- Included from
- Pro
- Z.ai
GLM 5.3 (Huawei Cloud)
A code model in the Theus catalogue.
- Output usage
- ×1TT / token
- Context
- 1M
- Included from
- Pro
- OpenAI
GPT-5.3-Codex
A multimodal model in the Theus catalogue.
- Output usage
- ×2.8TT / token
- Context
- 400K
- Included from
- Max 10x
- OpenAI
GPT-5.4
A general model in the Theus catalogue.
- Output usage
- ×4.5TT / token
- Context
- 1.1M
- Included from
- Pro
- OpenAI
GPT-5.6 Terra
A general model in the Theus catalogue.
- Output usage
- ×3TT / token
- Context
- 400K
- Included from
- Pro
Explore the quality and usage map
Theus Ranking · the map
Quality against weight.Every dot, one model.
Higher means more quality; further left, less weight per token. Tap a cell and the list below shows only its models.
Best for less: 5 level-4 models weigh ×1.
DeepSeek-V4 Pro · GLM 5.3 · GLM 5.3 (Huawei Cloud) · MiniMax M3 · Qwen3.7 Max
Map of 184 text models across 5 quality levels and per-token weight bands.
- Included from Free or Pro
- From Max
- From Max 10x
15 image, video, voice and embedding models are not on the map: they are not measured by level or by token. They are in the list.
Coming soon1
Announced, but not served yet: they do not count in the figure.
- Theus Local PrivadoTheus
How it is ordered
A ranking of our own.With the rules in plain sight.
Theus Quality
Every text model has a level from 1 to 5 set by Theus. Models on the same level share their position: no tie-breaks are invented.
Weight
How many TT each output token takes from your quota; input weighs less. ×1 is the reference. Image, video and voice are measured separately.
Plan
The first plan that includes the model. Each plan includes everything in the one below and adds the models that cost more.
Usage in Theus
In the model picker of your session —CLI, Desktop and chat— you can also order by usage in Theus. Usage figures are not published here.
Levels reviewed on 17 Oct 2026 · List taken from the gateway on 23 Oct 2026
See each model's weight →The catalogue is tiered:each plan adds its own step.
Free reaches 4 models and each tier unlocks the one below it. If you do not choose, TITAN answers — the autopilot that picks the model per task.