OpenAI compatible API · Attested · Public status

Qwen3 235B A22B Instruct 2507 vs Z.ai: GLM 4.6

Qwen3 235B A22B Instruct 2507 vs Z.ai: GLM 4.6: compare current API pricing, context, provider routes, privacy, p50 latency, and OpenAI-compatible access.

Verify gateway

Onebase URL to migrate

100sof models and routes

0prompt or output logs. Always.

Compare routesProviders, price, context, and policy posture in one view.

Use auto when uptime mattersKeep a primary model and let fallback handle provider failures.

Same API shapeUse the OpenAI client and set the model you want.

Practical read

Monthly evidence

Qwen3 235B A22B Instruct 2507 has the lower cheapest prompt+completion route on TrustedRouter. Qwen3 235B A22B Instruct 2507 has more provider fallback routes, while Z.ai: GLM 4.6 has the larger context window. Current TrustedRouter probes show Qwen3 235B A22B Instruct 2507 with the lower p50 TTFT.

$0.6752/1MQwen3 235B A22B Instruct 2507 cheapest route

$2.2999/1MZ.ai: GLM 4.6 cheapest route

1326 msQwen3 235B A22B Instruct 2507 fastest measured p50 TTFT via scaleway

1695 msZ.ai: GLM 4.6 fastest measured p50 TTFT via io-net

Prices and live route measurements can change. Use the monthly reports when you need a stable evidence window, then run a small task-specific eval before choosing a production default.

Qwen3 235B A22B Instruct 2507Z.ai: GLM 4.6

Model idqwen/qwen3-235b-a22b-instruct-2507z-ai/glm-4.6

Publisher Novita AI Z.AI

AI IQ — —

Context131,072 tokens204,800 tokens

Credits provider routes65

Cheapest route$0.6752/1M$2.2999/1M

Cached input$0.024265/1M$0.0844/1M to $0.28274/1M

Privacy postureprovider posture varieshas ZDR route

Fastest measured p50 TTFT1326 ms via scaleway1695 ms via io-net

Highest measured throughputnot enough datanot enough data

Recent route uptime range100.00%100.00%

Modes chat chat

Qwen3 235B A22B Instruct 2507 routes

Overview6 Credits routes
PricingInput, cached input, and output rates
ProvidersAll serving providers

Z.ai: GLM 4.6 routes

Overview5 Credits routes
PricingInput, cached input, and output rates
ProvidersAll serving providers

Related comparisons

Browse every comparison

Qwen3 235B A22B Instruct 2507 vs Z.ai: GLM 5.237 combined provider routes
Z.ai: GLM 4.6 vs Z.ai: GLM 5.236 combined provider routes
MoonshotAI: Kimi K3 vs Qwen3 235B A22B Instruct 250729 combined provider routes
DeepSeek: DeepSeek V4 Flash 0731 vs Qwen3 235B A22B Instruct 250728 combined provider routes
MoonshotAI: Kimi K3 vs Z.ai: GLM 4.628 combined provider routes
Qwen3 235B A22B Instruct 2507 vs Z.ai: GLM 5.328 combined provider routes
DeepSeek: DeepSeek V4 Flash 0731 vs Z.ai: GLM 4.627 combined provider routes
Z.ai: GLM 4.6 vs Z.ai: GLM 5.327 combined provider routes

Production choice

Pick a default model. Keep fallback enabled.

TrustedRouter is useful when you know the model you want, but still need provider rollover, budget limits, usage records, and a prompt path you can verify.

OpenAI clientPython

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

response = client.chat.completions.create(
    model="qwen/qwen3-235b-a22b-instruct-2507",
    messages=messages,
)

Questions

Which should I use, Qwen3 235B A22B Instruct 2507 or Z.ai: GLM 4.6?

Is Qwen3 235B A22B Instruct 2507 or Z.ai: GLM 4.6 cheaper?

The current cheapest TrustedRouter route is $0.6752/1M for Qwen3 235B A22B Instruct 2507 and $2.2999/1M for Z.ai: GLM 4.6. The comparison uses current catalog prices and updates as provider pricing changes.

Is Qwen3 235B A22B Instruct 2507 or Z.ai: GLM 4.6 faster?

Current measured p50 time to first token is 1326 ms for Qwen3 235B A22B Instruct 2507 and 1695 ms for Z.ai: GLM 4.6. These are routed probe measurements, not vendor-advertised speeds, and update as new samples arrive.

Can I test Qwen3 235B A22B Instruct 2507 and Z.ai: GLM 4.6 with the same API?

Yes. Use the same OpenAI-compatible TrustedRouter base URL and API key, then change only the model id between qwen/qwen3-235b-a22b-instruct-2507 and z-ai/glm-4.6. This makes side-by-side evals possible without maintaining two provider integrations.