Discover models
Practical starting points from real SwitchGate usage, live production performance, and independent benchmarks.
Today's frontier
Claude Opus takes the frontier lead
Anthropic edges OpenAI by two points on the intelligence index. All figures are placeholder data until the live catalog ships.
Intelligence (0–100)
Frontier leaderboard
- 1Claude Opus 4.1Anthropic63
- 2GPT-5.2OpenAI61
- 3Gemini 3 ProGoogle58
- 4Grok 4.1xAI56
- 5DeepSeek V3.2DeepSeek52
#1 on SwitchGate
deepseek/deepseek-v3.2
10Ttokens/wk
Smartest open model
groq/kimi-k2-instruct
60intelligence
Best value
google/gemini-3-flash
$0.15/M input tokens
Fastest
groq/llama-4-scout
812tok/s
Featured models
100+ active models, one endpoint
Weekly token volume through the gateway, at provider list price, pass-through.
Value leaders
View more →The most capability per dollar of list price.
| Model | Intel pctl | Coding pctl |
|---|---|---|
| gemini-3-flash | 53 | 60 |
| deepseek-v3.2 | 80 | 84 |
| mistral-small-4 | 64 | 72 |
| gpt-5-mini | 81 | 88 |
| qwen3-235b | 55 | 69 |
Fastest models
View more →Highest median output speed across providers.
| Model | Throughput | TTFT p50 |
|---|---|---|
| llama-4-scout | 812 tok/s | 263ms |
| llama-4-maverick | 359 tok/s | 313ms |
| gemini-3-flash | 299 tok/s | 268ms |
| grok-4-fast | 289 tok/s | 261ms |
| nova-lite | 272 tok/s | 277ms |
Full catalog
Every model, one endpoint
Search the catalog below; every ID is callable through one OpenAI-compatible endpoint at provider list price.
No models match your search.
Placeholder data for design purposes; the live catalog ships with the API.
One endpoint for all of them
Set base_url to https://api.switchgate.ai/v1 and call any model ID above.