Hy3 vs Hy4 preview
At a Glance
| Compare |  Hy3Tencent | Hy4 previewTencent |
|---|---|---|
| Intelligence, Cost, and Efficiency | ||
| IntelligenceHigher is better · MM Intelligence v4.4 | #25 of 5750.1 score · 1/3 sources · single-source provisional · missing LiveBench, ARC-AGI-2 · full-core range 16.7–83.4 | UnrankedNot in the 57-model eligible cohort |
| Pricing and Limits | ||
| Input priceFrom · USD / 1M tokens | $0.13Deepinfra ↗ · Oct 1, 2026 | $0.834Deepinfra ↗ · Oct 19, 2026 |
| Output priceFrom · USD / 1M tokens | $0.528Openrouter ↗ · Oct 19, 2026 | $2.501Deepinfra ↗ · Oct 19, 2026 |
| Context windowMaximum documented tokens | 262K | 1,000K |
| Model facts checked | Aug 28, 2026View model evidence → | Sep 2, 2026View model evidence → |
Token prices are the lowest available sourced USD rates; input and output may use different providers. Cost ranking estimates output spend on LiveBench, not a full request bill. Ranking methodology →
Available Benchmarks
| Benchmark | Hy3 | Hy4 preview |
|---|---|---|
| LMArena Agent Arenaagent-2026-10-16-a64ca3b12b6e · outcome_score · leader | -6.0092% of row best · score · Hy3; 95% CI [-7.50774599, -4.49280945]; sessions 31296; observations 1149668; rank 41 | 2.68100% of row best · score · Hy4 preview; 95% CI [2.00929922, 3.35139547]; sessions 84828; observations 8314177; rank 20 |
| Overall ResultCounted from the protocol-matched rows above | 0 benchmark winsNo overall winner | 1 benchmark winNo overall winner |
Third-party benchmark Only like-for-like primary-publisher results are shown; raw scores, relative scores, configuration, and token spend remain visible.
Side-by-Side Facts
| Field | Hy3 | Hy4 preview |
|---|---|---|
| Developer | Tencent | Tencent |
| Family | Hy3 | Hy4 |
| Model | Hy3 | Hy4 preview |
| Version | Hy3 | Hy4 preview |
| Lifecycle | active | preview |
| Released | Unknown | 2026-08-28 |
| Knowledge cutoff | Unknown | Unknown |
| Input modalities | Text | Text |
| Output modalities | Text | Text |
| Context window | 262K | 1,000K |
| Total parameters | 298.8B | 770B |
| Active parameters | 21B | 49B |
| License | apache-2.0 | apache-2.0 |
| Open weights | Yes | Yes |
| API available | Yes | Yes |
| Self-hostable | Yes | Yes |
| Provider access | Deepinfra (Standard), Hugging Face (Standard), Openrouter (Standard) | Deepinfra (Standard), Openrouter (Standard) |
| Capabilities | chat, generation, tools | chat, generation, reasoning, tools |
Hy3 Capabilities
Hy4 preview Capabilities
Related Comparisons
All text comparisons →| A | Pair | B | Context |
|---|---|---|---|
 MiniMax-M3MiniMax | vs | Hy3Tencent | cross-developer peerstext |
| vs | Hy3Tencent | cross-developer peerstext | |
 ERNIE X1.1Baidu | vs | Hy3Tencent | cross-developer peerstext |
 Antigravity AgentGoogle DeepMind | vs | Hy3Tencent | cross-developer peerstext |
| vs | Hy3Tencent | cross-developer peerstext | |
| vs | Hy3Tencent | cross-developer peerstext | |
MiniMax-M2.5MiniMax | vs | Hy3Tencent | cross-developer peerstext |
MiniMax-M2.7MiniMax | vs | Hy3Tencent | cross-developer peerstext |
 Devstral-2-123B-Instruct-2512Mistral AI | vs | Hy3Tencent | cross-developer peerstext |
| vs | Hy3Tencent | cross-developer peerstext | |
| vs | Hy3Tencent | cross-developer peerstext |
Primary Evidence
Sources and Freshness
Questions
Hy3 vs Hy4 preview FAQs
Is Hy3 or Hy4 preview better for coding?+
This comparison does not currently contain a protocol-matched coding benchmark for both Hy3 and Hy4 preview, so Model Markets cannot name a coding leader from pricing, context size, or capability labels alone.
Which is cheaper, Hy3 or Hy4 preview?+
Hy3 is $0.13 and Hy4 preview is $0.834 per million tokens, so Hy3 is cheaper on this metric. Hy3 is $0.528 and Hy4 preview is $2.501 per million tokens, so Hy3 is cheaper on this metric.
Which has a larger context window, Hy3 or Hy4 preview?+
Hy4 preview has the larger sourced context window. Hy3 supports 262,144 tokens and Hy4 preview supports 1,000,000 tokens.
Which performs better in benchmarks, Hy3 or Hy4 preview?+
There is no overall benchmark winner: At least two independently verified, protocol-matched benchmarks are required for an overall winner.
Can Hy3 or Hy4 preview be self-hosted?+
Both models have the same sourced self-hosting status: supported. Hy3 is open weight; Hy4 preview is open weight.
Can Hy3 and Hy4 preview understand images?+
Hy3 is not documented with image input; Hy4 preview is not documented with image input. This reflects supported input modalities, not vision quality.
Which can generate longer answers, Hy3 or Hy4 preview?+
Neither model has a sourced maximum output in this comparison, so a larger limit cannot be determined.
Do Hy3 and Hy4 preview support reasoning and tool use?+
Hy3: tool calling. Hy4 preview: reasoning and tool calling. Feature support does not establish relative quality.
Which is available from more inference providers, Hy3 or Hy4 preview?+
Hy3 has 3 sourced provider routes; Hy4 preview has 2, so Hy3 has broader tracked availability.
Which offers better value, Hy3 or Hy4 preview?+
There is no universal value winner. Compare the input and output prices above with the matched benchmark result for your workload: cheaper tokens can be offset by different quality, token usage, latency, or provider availability.