GLM 5.3 Flash on QuickSilver Pro
GLM 5.3 Flash is Z.ai's fast open-weights sibling of GLM 5.3 — the same 2026 reasoning family distilled to workhorse pricing, with the full 1M-token context window and 131K max output. On QuickSilver Pro it is $0.07 input / $0.05 output per million tokens, at parity with the OpenRouter endpoint selected by this route, with no markup or discount. Like GLM 5.3, reasoning is always on for this endpoint and cannot be disabled — control depth with `reasoning_effort` instead.
At a glance
High-volume agents and coding at workhorse pricing — the fast open-weights member of Z.ai's newest reasoning family, with a 1M-token context window.
Pricing comparison ($/1M tokens)
| Provider | Input | Output | vs QSP |
|---|---|---|---|
| QuickSilver Pro | $0.07 | $0.05 | — |
| OpenRouter (z-ai/glm-5.3-flash) | $0.07 | $0.05 | same |
| OpenAI (GPT-4o mini) | $0.15 | $0.60 | 92% lower |
When to use
Reach for GLM 5.3 Flash when you want the GLM 5.3 family's reasoning-and-tools behavior at a fraction of the price: high-volume agent loops, background codegen, summarization over huge working sets in its 1M-token context, and batch pipelines where per-call cost dominates. At $0.07 / $0.05 per million tokens it competes with the cheapest tier of the catalog while keeping a 1M-token window — a combination the budget models don't offer. Reasoning is always on; use `reasoning_effort` to trade depth against latency and cost.
When to use something else
For the hardest repo-scale engineering and long-horizon plans, GLM 5.3 is the stronger family pick — same window, deeper model. Reasoning is mandatory here too, so it spends thinking tokens even on trivial turns; if you want a cheap model that replies directly, DeepSeek V4 Flash ($0.06/$0.18) answers without the overhead. If you have prompts tuned on GLM 5.2 or GLM 5.3, A/B before switching down — traces and tool-call cadence differ across the family.
Quickstart (curl)
curl https://api.quicksilverpro.io/v1/chat/completions \
-H "Authorization: Bearer $QSP_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "glm-5.3-flash",
"messages": [{"role": "user", "content": "Hello!"}]
}'OpenAI-compatible. One-line migration via base_url.
FAQ
Reasoning is always on — it is mandatory on this endpoint and cannot be disabled (a request that tries to turn it off is rejected upstream). Use `reasoning_effort` to trade depth against latency and cost. Reasoning tokens bill as output tokens and are drawn from your `max_tokens` budget before the visible answer, so set `max_tokens` generously or the reply can come back truncated. If you want a GLM without always-on thinking, use GLM 5.2.
Same family, different weight class: GLM 5.3 Flash is the fast, open-weights variant released 2026-08-26, priced at $0.07 input / $0.05 output per million tokens versus $0.60 / $2.20 for GLM 5.3 — about 8.6x lower on input. Both share the 1M-token context window, the 131K max-output limit, and the always-on reasoning contract. Use Flash for volume; step up to GLM 5.3 when task difficulty, not throughput, is the bottleneck.
Yes — GLM 5.3 Flash is an OpenAI-compatible chat completions endpoint on QuickSilver Pro (the Responses API works too). Set base_url=https://api.quicksilverpro.io/v1, paste your QSP key, and use model="glm-5.3-flash". Streaming, tool calling (`tool_choice: "auto"` — this endpoint rejects forcing a specific function), and usage.cost accounting all work. `response_format: json_schema` is not enforced on this endpoint yet; if you need strict structured output, use GLM 5.2.
$0.07 input / $0.05 output and $0.03 cache read per million tokens, matching the route-selected OpenRouter endpoint (inference-net fp4) with no markup or discount. The route uses its pinned provider order or the cheapest eligible endpoint under sort: price and its quantization filters; the /models headline can differ. More expensive fallback hosts do not change your price. Use base_url=https://api.quicksilverpro.io/v1 and model="glm-5.3-flash".