Hy3
by Tencent · tencent/hy3 · #73 cheapest of 325 paid models
Prices updated Jul 22, 2026, 1:31 AM 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
$0.14
per 1M tokens
Output
$0.58
per 1M tokens
Blended (3:1)
$0.25
3 input : 1 output
Cache read
$0.035
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% | $0.25 | $0.0149 | — |
| 50% | $0.211 | $0.00961 | −35% |
| 90% | $0.179 | $0.00541 | −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 | $0.000244 | $0.244 |
| Document summary | 8,000 | 1,000 | $0.0017 | $1.70 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.00536 | $5.36 |
| Long-context analysis | 150,000 | 5,000 | $0.0239 | $23.90 |
Performance
Intelligence Index
41.2
AA composite quality
Coding Index
58.8
Math Index
—
Agentic Index
—
Output speed
61 tok/s
median
Time to first token
1.69s
median
Time to first answer
34.64s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| GPQA Diamond | 89.7% |
| Humanity's Last Exam | 31.6% |
| SciCode | 47.6% |
| AA-LCR (Long Context Reasoning) | 66.7% |
| Terminal-Bench 2.1 | 64.4% |
| τ³-Bench Banking | 20.8% |
Design Arena Elo
| Category | Elo |
|---|---|
| Website | 1,228 |
| UI components | 1,241 |
| Data viz | 1,186 |
| 3D | 1,269 |
| Game dev | 1,188 |
| Code categories | 1,240 |
Specs
Context window
262K
tokens
Max output
262K
tokens
Input modalities
text
Tokenizer
Other