- 1M context
- Multimodal
- Reasoning
- Responses API
- NEW · Nov 16
Hy4 Flash on QuickSilver Pro
Hy4 Flash is the lowest-priced Tencent option on QuickSilver Pro, below Hy4. Text and image input, optional reasoning and a 1M-token context at $0.0385 input / $0.1546 output per 1M tokens.
- 1M context
- 66K max output
- Vision
- Tools
- Reasoning · on by default
- Streaming
- JSON Schema output
You pay · input / output per 1M
$0.0385 / $0.1546
Tencent list source: huggingface.co/tencent/Hy4-flash
At a glance
- Context
- 1M tokens
- Input / 1M
- $0.0385
- Output / 1M
- $0.1546
- Thinks by default
- Yes
- Max output
- 66K
- Latency · live
- —
Low-cost text, image analysis and automatic tool calls with optional reasoning.
When is Hy4 Flash the right call?
Curated from catalog capabilities and published prices, not a benchmark.
When to use
Choose Hy4 Flash for budget-sensitive text and image workflows, automatic tools and strict JSON Schema. These features passed small upstream probes; evaluate your own workload before scaling.
- Long documents and codebases, up to 1M tokens of context
- Agents that call tools, with strict JSON Schema output
- Prompts that include images
- Multi-step reasoning, on by default
When to use something else
Choose another model for forced named or required tool calls. Temperature is supported; top_p is dropped. No coding benchmark or long-context quality claim is made from these small probes.
Quickstart (curl)
Get an API key →curl https://api.quicksilverpro.io/v1/chat/completions \
-H "Authorization: Bearer $QSP_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "hy4-flash",
"messages": [{"role": "user", "content": "Hello!"}]
}'import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.quicksilverpro.io/v1",
api_key=os.environ["QSP_KEY"],
)
response = client.chat.completions.create(
model="hy4-flash",
messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.quicksilverpro.io/v1",
apiKey: process.env.QSP_KEY,
});
const response = await client.chat.completions.create({
model: "hy4-flash",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(response.choices[0].message.content);curl https://api.quicksilverpro.io/v1/responses \
-H "Authorization: Bearer $QSP_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "hy4-flash",
"input": "Hello!"
}'import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.quicksilverpro.io/v1",
api_key=os.environ["QSP_KEY"],
)
response = client.chat.completions.create(
model="hy4-flash",
messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)OpenAI-compatible. One-line migration via base_url.
Gateway behavior
Compatibility notes for this model.
top_p- top_p is dropped; temperature is retained.
prompt- Prompts up to 1M tokens.
max_tokens- Up to 66K output tokens.
- /v1/chat/completions: Supported
- /v1/responses: Supported
- streaming: Supported
- tools: Supported
- image input: Supported
- JSON Schema: Supported
- reasoning: Supported
Where the price comes from
- Provider reference
- No provider list price is compared for this model.huggingface.co/tencent/Hy4-flash
- Catalog verified
- Prompt cache · not compared
- $0.00165 read
usage.cost = prompt_tokens × $0.0385/1M + completion_tokens × $0.1546/1M
FAQ
Common questions
Set base_url=https://api.quicksilverpro.io/v1 and model="hy4-flash". The gateway supports Chat Completions and Responses. The listed limits are 1M context tokens and 66K output tokens; full-context generation was not tested.
Reasoning is on by default. Small upstream probes returned zero reasoning tokens with reasoning.enabled=false or reasoning.effort="none"; enabled=true and effort="low" produced reasoning. Reasoning tokens are included in completion tokens and billed once as output.
$0.0385 input / $0.1546 output per 1M tokens; cached input $0.00165. Current pricing reflects a temporary promotion and may change. Repeated long-prompt probes reported cache hits. No cache-write charge was listed or observed.