MODEL PROOF
Hy3
BETTER VALUE OPTION
A stored alternative has an equal-or-higher measured score and an equal-or-lower price.
Standard-window rate shown; discount 16:00–24:00 UTC every day: $0.083 in · $0.330 out. · Serving precision differs between offers.
DeepSeek V4.1 Flash · Δ score 21.8 points · 21% lower measured price · $0.182 / 1M tokens
Gemma 4 31B is a further stored option with these losses: 128K → 16K max output
- Independent LMArena score
- 1,440.7 ± 6.5 · 10,476 votes
- Context window
- 262.14K
- Maximum output
- 128K
- Input modalities
- text
- Output modalities
- text
- Published input price
- $0.132 / 1M tokens
- Published output price
- $0.528 / 1M tokens
- Pricing kind
- fixed
Inspect complete billing conditions and endpoint terms below.
Price from Tencent: the cheapest offer that is serving, at standard delivery and at the maker’s precision or better, at its standard rate.
Cheaper right now: $0.230/M at DeepInfra — 4-bit It does not pass the like-for-like test, so it never sets a rank.
Every price dimension · USD per million tokens
| Input | $0.132/M |
|---|---|
| Output | $0.528/M |
| Cached input | $0.033/M |
| Cache write | Unknown/M |
| Cache write 1h | Unknown/M |
| Reasoning | Unknown/M |
Price varies by time of day. The headline is the standard rate, $0.132 in · $0.528 out. Discount window 16:00–24:00 UTC every day: $0.083 in · $0.330 out.The discount is 38% off input.The windows come from the tencent/fp8 endpoint in the upstream endpoints feed, not from the time this page was built.
Who sells it
The same weights, different shops. Cheapest is not like-for-like when serving precision differs.
Headline cheapest is a lower precision. Like-for-like at the best declared precision is $0.140/M from GMICloud.
6 provider offers · rates, limits and conditions
Rates are USD per million tokens. Endpoint terms can differ even when the quoted price and precision match. A listed parameter is a provider declaration, not a task-success test.
| Seller | Input | Output | Precision | Uptime (1d) |
|---|---|---|---|---|
| DeepInfradeepinfra/fp4 Endpoint terms
| $0.130 | $0.530 | fp4 | 99.81% |
| Tencenttencent/fp8 Endpoint terms
| $0.132 | $0.528 | fp8 | 99.81% |
| GMICloudgmicloud/bf16 Endpoint terms
| $0.140 | $0.580 | bf16 | 99.91% |
| Novitanovita Endpoint terms
| $0.140 | $0.580 | undeclared | 99.52% |
| Phalaphala Endpoint terms
| $0.150 | $0.640 | undeclared | 99.81% |
| AtlasCloudatlas-cloud/fp8 Endpoint terms
| $0.200 | $0.800 | fp8 | 99.92% |
Independent scores
| LMArena Elo | 1440.7default |
|---|
LMArena Elo under CC BY 4.0. Absence of another board is not a score of zero.
Best rankings by task
- 3D #50 1184
- Code categories #60 1181
- Websites #65 1189
- UI components #67 1170
- Game development #71 1149
- Data visualisation #84 1132
Per-task Elo and rank from Design Arena, via OpenRouter’s model feed. The rank is the feed’s, and it counts models OpenRouter does not list.
Capability
| Context window | 262K |
|---|---|
| Max output | 128K |
| Input modes | text |
| Tool use | yes |
| Reasoning | optional |
| Open weights | Unknown |
Provenance
| Price source | openrouter.ai |
|---|---|
| Fetched | 2026-10-06 |
| Quality data | verified |
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:...
Questions this page answers
What does Hy3 cost?
$0.132/M in, $0.528/M out in this catalogue, as of the fetch date on this page. That is the API row, not a subscription.
Does Hy3 have an independent quality score?
Hy3 has an LMArena score in this catalogue.
What beats Hy3?
DeepSeek V4.1 Flash has an equal-or-higher measured score and an equal-or-lower price.
Does this page use Artificial Analysis scores for Hy3?
No. Artificial Analysis figures are not published here. Quality on this page is LMArena Elo where a score exists; otherwise the row is unrated.
Is the cheapest Hy3 endpoint the same product?
Headline cheapest is a lower precision. Like-for-like at the best declared precision is $0.140/M from GMICloud.
MODEL MONUMENT
Save or share this proof
Badge
The verdict as one SVG, for a README or a docs page. It states this model's dominance status at the balanced workload on the LMArena lens, with the date it was computed, and it is rebuilt with the catalogue — so it changes when the verdict changes, including to one you would rather it did not.
[](https://undominated.ai/models/tencent__hy3/)<a href="https://undominated.ai/models/tencent__hy3/"><img src="https://undominated.ai/badge/tencent__hy3.svg" alt="Hy3 — Undominated.ai dominance verdict" height="36"></a>Direct file: https://undominated.ai/badge/tencent__hy3.svg — a static SVG, written by scripts/build-badges.mjs on every build, so the copy you embed is never
older than the last deploy.
