OpenAI compatible API · Attested · Public status

Latest AI Model APIs: Kimi K3, GLM 5.2, DeepSeek V4

Call Kimi K3, GLM 5.2, DeepSeek V4, and Gemini 3.6 Flash through one OpenAI-compatible API with published pricing, privacy labels, and fallback.

Verify gateway

Onebase URL to migrate

100sof models and routes

0prompt or output logs. Always.

New model APIs

Try the models developers are evaluating now.

Kimi K3, GLM 5.2, DeepSeek V4, and Gemini 3.6 Flash share one OpenAI-compatible endpoint. Create one key, keep one SDK, and compare them on your own prompts.

Compare the catalog

New email signups receive starter credit. No card required.

Switch the model stringOpenAI SDK

from openai import OpenAI

client = OpenAI(
    base_url="https://api.trustedrouter.com/v1",
    api_key="sk-tr-v1-...",
)

response = client.chat.completions.create(
    model="moonshotai/kimi-k3",
    messages=[{"role": "user", "content": "Test this task"}],
)

# Try z-ai/glm-5.2 or deepseek/deepseek-v4-pro next.

Kimi K3

Long-context reasoning

Compare every currently available Kimi K3 provider route, price, privacy label, and live performance.

Open Kimi K3 GLM 5.2

Open-weight frontier route

Call GLM 5.2 through several providers and choose a route by cost, speed, region, or privacy.

Open GLM 5.2 DeepSeek V4

Cost-efficient reasoning

Use V4 Pro for quality or compare the Flash routes when throughput and price matter more.

Open DeepSeek V4 Gemini 3.6 Flash

Fast multimodal work

Compare Google AI Studio and Vertex routes without changing your application integration.

Open Gemini 3.6 Flash

One integration

Evaluate on your task, then ship the winner.

Same request shape. Keep one OpenAI-compatible client and switch the model ID instead of maintaining separate provider SDKs.

Route evidence. Model pages publish providers, pricing, context, measured performance, retention posture, and verified confidential-compute status.

Provider rollover. Models served by multiple providers can move to a healthy authorized route when the first provider returns a retryable failure.

Privacy with proof. TrustedRouter keeps no prompt or output logs, always. Upstream guarantees differ and remain explicit on every route.

One key. Hundreds of models. Run Kimi K3, GLM 5.2, DeepSeek V4, or Gemini 3.6 Flash with the SDK you already use.

No card required.

Live catalog evidence

Current routes, prices, privacy, and measured performance.

Catalog facts come from the routes currently configured in TrustedRouter. Performance uses the same cached metadata snapshot as the public leaderboard. Prompts and outputs are not part of these measurements.

550public models

85providers

1677configured routes

297ZDR routes

32provider E2EE routes

4021recent availability samples

Model	Providers	Context	Input	Output	Privacy	Measured route
Anthropic: Claude Opus 4.8`anthropic/claude-opus-4.8`	3 routes	1,000,000	$5.275/1M	$26.375/1M	varies 5 cited scores	4362 ms TTFT anthropic · 62 tok/s · 100.00% available · n=17
OpenAI: GPT-5.5`openai/gpt-5.5`	4 routes	1,050,000	$5.275/1M	$31.65/1M	ZDR 3 cited scores	Warming up
Google: Gemini 3.5 Flash`google/gemini-3.5-flash`	+1 7 routes	1,048,576	$1.5825/1M	$9.495/1M	ZDR	2763 ms TTFT google-vertex · 78 tok/s · 100.00% available · n=7
MoonshotAI: Kimi K2.7 Code`moonshotai/kimi-k2.7-code`	+11 22 routes	262,144	$0.70685/1M to $1.07188/1M	$3.587/1M to $4.77915/1M	ZDR 5 cited scores	1616 ms TTFT fireworks · 48 tok/s · 100.00% available · n=33
Z.ai: GLM 5.2`z-ai/glm-5.2`	+30 53 routes	1,048,576	$0.7174/1M to $2.325744/1M	$1.5825/1M to $7.106438/1M	E2EE 4 cited scores	1796 ms TTFT engy · 29 tok/s · 100.00% available · n=161
MiniMax: MiniMax M3`minimax/minimax-m3`	+8 20 routes	1,048,576	$0.24265/1M to $0.633/1M	$1.0128/1M to $2.532/1M	ZDR 4 cited scores	1774 ms TTFT minimax · 98 tok/s · 100.00% available · n=19

AnthropicPolicy varies 10 models 3564 ms p50 · n=121

GMI CloudPolicy varies 67 models 2925 ms p50 · n=47

OpenAIZDR on prepaid 44 models 1056 ms p50 · n=500

Atlas CloudPolicy varies 75 models 1961 ms p50 · n=22

Google AI StudioPolicy varies 12 models 1853 ms p50 · n=499

Google Vertex AIZDR on prepaid 10 models 2763 ms p50 · n=28

Lightning AIPolicy varies 36 models 2195 ms p50 · n=20

KimiPolicy varies 11 models 1759 ms p50 · n=45

Browse every model Review provider policies Open the full leaderboardSnapshot 2026-08-25T07:59:45.277Z

Questions

Which new AI models are available through TrustedRouter?

Current highlighted routes include Kimi K3, GLM 5.2, DeepSeek V4 Pro and Flash, and Gemini 3.6 Flash. The live models catalog is the source of truth and lists hundreds of additional routes, their providers, context limits, prices, and privacy posture.

Do I need a separate account for every model provider?

No. Prepaid TrustedRouter credits provide one key and one OpenAI-compatible base URL across supported providers. Bring-your-own provider keys are also available when you prefer to keep a direct provider relationship.

Can I switch models without changing SDKs?

Yes. Keep the OpenAI SDK and TrustedRouter base URL, then change only the model string. The same key can call Kimi K3, GLM 5.2, DeepSeek V4, Gemini 3.6 Flash, and the rest of the live catalog.

Does every model route have the same privacy guarantee?

No. TrustedRouter itself keeps no prompt or output logs, always, while upstream provider behavior differs. Every route publishes its provider, Zero-Data-Retention status, and verified confidential-compute status so you can choose deliberately.