The AI arena is free today

Open Superagent
TencentReleased on Nov 16, 2026

Hy4 flash: Benchmarks, Pricing & Context Window

Hy4 flash is a language model from Tencent, released in November 2026, with multimodal input, a 960K-token context window, and pricing from $0.070/M input, $0.003/M cached input, $0.281/M output.

Hy4 flash is Tencent Hy's most cost-effective Hy model, announced on 2026-11-16 (API id hy4-flash). It has 224 billion total parameters, including 64 billion Engram parameters, with 12 billion active parameters per token. Like Hy4, it is

Input
TextImage
Output
Text

Hy4 flash benchmarks

Capability tiers

Standing within each category, adjusted for leaderboard depth.

Real tasks performance

High-confidence performance for Hy4 flash across real-world prompt categories. Only 95% intervals at most 4 points wide are shown.

Performance by conversation depth

How Hy4 flash holds up as conversations get longer.

Quality Tracker

Hy4 flash Performance Across Datasets

Scores sourced from the model's scorecard, paper, or official blog posts

LLM Stats Logollm-stats.com - Mon Nov 16 2026
Notice missing or incorrect data?

Hy4 flash pricing

Providers

Hy4 flash starts at $0.0700 per million input tokens and $0.281 per million output tokens via Tencent Cloud. Reused prompt prefixes cost $0.0030 per million cached input tokens.

ProviderInput $/MCached input $/MOutput $/MContext in / outTTFT p95 sOutput p5 c/sModalities in / out
Tencent Cloud logoTencent Cloud
$0.0700$0.0030$0.281960.0K/64.0K
—
—
/

Cached input is the discounted price for prompt tokens served from a provider cache. TTFT is time to first token. Output is characters per second; p5 is the sustained floor exceeded by 95% of observed requests.

Hy4 flash context window

Input and output token limits for Hy4 flash, plus how it ranks on long-context understanding.

InputOutput
960Ktokens
64Ktokens
≈ 1.4k pages of text
960K
8K128K1M

Try now

huggle
Hy4 flashin Huggle

Make it with
Hy4 flash.

Hy4 flash

Hy4 flash latency

Hy4 flash time to first token, sustained output throughput, and failed-request rate from live model usage over the trailing 7 days.

Hy4 flash examples

Recent arena outputs from Hy4 flash, picked from the highest-ranked matchups.

Hy4 flash license

Hy4 flash has 224.0B parameters.

Parameters
224.0B

Hy4 flash resources

Official sources for Hy4 flash: provider documentation, official launch post.

Hy4 flash vs other models

The most-compared alternatives to Hy4 flash are GPT-5.5 Instant, Qwen3.6-27B, LongCat-2.5. Open any pair side-by-side for benchmarks, pricing, context, and latency.

Models like Hy4 flash

Models ranked just above and below Hy4 flash by LLM Stats score.

 

GPT-5.5 Instant

Score pending
 

Qwen3.6-27B

Score pending
 

LongCat-2.5

Score pending
 

o3

Score pending
 

Claude Sonnet 4.6

Score pending
 

GLM-5.3-Flash

Score pending

FAQ

Common questions about Hy4 flash.

When was Hy4 flash released?

Hy4 flash was released on November 16, 2026 by Tencent. This is the official Hy4 flash release date tracked on LLM Stats.

How much does Hy4 flash cost?

Hy4 flash pricing starts at $0.07 per million input tokens, $0.00 per million cached input tokens, $0.28 per million output tokens via Tencent Cloud, the lowest price among tracked providers.

How big is Hy4 flash?

Hy4 flash has 224 billion parameters.

Who created Hy4 flash?

Hy4 flash was created by Tencent.

Is Hy4 flash multimodal?

Yes, Hy4 flash is multimodal and can accept both text and images as input.

Where can I use Hy4 flash?

Hy4 flash is available through 1 provider including Tencent Cloud.

Where is the Hy4 flash paper or technical report?

Hy4 flash has a paper or technical report available at https://hy.tencent.ai/research/hy4-flash. Use that source for architecture, training, release and evaluation details.

What models should I compare Hy4 flash against?

Common Hy4 flash comparisons include Hy4 flash vs GPT-5.5 Instant, Hy4 flash vs Qwen3.6-27B, Hy4 flash vs LongCat-2.5. Compare them side by side for benchmark scores, pricing, context window, latency and provider availability.