Hy3 preview
by Tencent · tencent/hy3-preview · #27 cheapest of 325 paid models
Prices updated Jul 22, 2026, 1:31 AM UTC · refreshed hourly
Hy3 preview is a high-efficiency Mixture-of-Experts model from Tencent designed for agentic workflows and production use. It supports configurable reasoning levels across disabled, low, and high modes, allowing it to...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$0.063
per 1M tokens
Output
$0.21
per 1M tokens
Blended (3:1)
$0.10
3 input : 1 output
Cache read
$0.021
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for Hy3 preview with prompt caching.
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $0.10 | $0.006636 | — |
| 50% | $0.084 | $0.004536 | −32% |
| 90% | $0.071 | $0.002856 | −57% |
*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates (available for OpenAI / Anthropic / Google — not flagged for this creator). Confirm on the vendor’s pricing page.
What a request costs
| Workload | Input tokens | Output tokens | Cost / request | Cost / 1K requests |
|---|---|---|---|---|
| Short chat message | 500 | 300 | $9.45e-5 | $0.0945 |
| Document summary | 8,000 | 1,000 | $0.000714 | $0.714 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.00231 | $2.31 |
| Long-context analysis | 150,000 | 5,000 | $0.0105 | $10.50 |
Performance
Intelligence Index
33.6
AA composite quality
Coding Index
58.8
Math Index
—
Agentic Index
30.7
Output speed
150 tok/s
median
Time to first token
1.68s
median
Time to first answer
15.05s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| GPQA Diamond | 86.7% |
| Humanity's Last Exam | 25.5% |
| SciCode | 41.2% |
| IFBench | 63.1% |
| AA-LCR (Long Context Reasoning) | 54.7% |
| Terminal-Bench Hard | 34.1% |
| τ²-Bench (Telecom) | 92.7% |
Specs
Context window
262K
tokens
Max output
—
tokens
Input modalities
text
Tokenizer
Other