Hy3
by Tencent · tencent/hy3 · #65 cheapest of 422 paid models
Prices updated Sep 19, 2026, 6:21 PM UTC · refreshed hourly
Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$0.083
per 1M tokens
Output
$0.33
per 1M tokens
Blended (3:1)
$0.144
3 input : 1 output
Cache read
$0.021
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for Hy3 with prompt caching.
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $0.144 | $0.008745 | — |
| 50% | $0.121 | $0.005651 | −35% |
| 90% | $0.103 | $0.003176 | −64% |
*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 | $0.00014 | $0.1402 |
| Document summary | 8,000 | 1,000 | $0.00099 | $0.99 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.003135 | $3.14 |
| Long-context analysis | 150,000 | 5,000 | $0.014 | $14.03 |
Performance
Intelligence Index
25.8
AA composite quality
Coding Index
58.8
Math Index
—
Agentic Index
—
Output speed
82 tok/s
median
Time to first token
1.70s
median
Time to first answer
26.09s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| GPQA Diamond | 89.7% |
| Humanity's Last Exam | 33.5% |
| SciCode | 48.6% |
| AA-LCR (Long Context Reasoning) | 79.0% |
| Terminal-Bench 2.1 | 64.4% |
| τ³-Bench Banking | 22.9% |
Design Arena Elo
| Category | Elo |
|---|---|
| Website | 1,194 |
| UI components | 1,178 |
| Data viz | 1,141 |
| 3D | 1,205 |
| Game dev | 1,160 |
| Code categories | 1,192 |
Specs
Context window
262K
tokens
Max output
128K
tokens
Input modalities
text
Tokenizer
Other