Tencent: hy3:free
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: a direct no-think mode by default, plus low and high chain-of-thought modes for complex math, coding, and multi-step problems. With a 256K context window, Hy3 targets long-horizon tasks, including improved coreference resolution, multi-turn constraint tracking, and stable tool-calling that generalizes across agent scaffoldings. Tencent positions it as a reliable, cost-effective option across coding, document processing, financial analysis, game development, and frontend design, with a strong emphasis on grounded, anti-hallucination behavior that answers when grounded and flags when evidence is missing rather than fabricating.
Preise
Verfügbarkeit
Leistung
Nutzung & Rang
Unterstützte Parameter
Alle Anbieter = jeder Upstream, der dieses Modell bereitstellt, unterstützt den Parameter. Einige Anbieter = hängt vom Upstream ab, der die Anfrage bearbeitet. Standard = der Wert, der gesendet wird, wenn nichts gesetzt ist.
| Parameter | Anbieter | Standard |
|---|---|---|
| frequency_penalty | Alle Anbieter | Standardmäßig nicht gesendet |
| include_reasoning | Alle Anbieter | - |
| logit_bias | Einige Anbieter | - |
| max_completion_tokens | Einige Anbieter | - |
| max_tokens | Alle Anbieter | - |
| min_p | Einige Anbieter | - |
| presence_penalty | Alle Anbieter | Standardmäßig nicht gesendet |
| reasoning | Alle Anbieter | - |
| reasoning_effort | Alle Anbieter | - |
| repetition_penalty | Alle Anbieter | Standardmäßig nicht gesendet |
| response_format | Einige Anbieter | - |
| seed | Alle Anbieter | - |
| stop | Alle Anbieter | - |
| structured_outputs | Alle Anbieter | - |
| temperature | Alle Anbieter | 0.9 |
| tool_choice | Alle Anbieter | - |
| tools | Alle Anbieter | - |
| top_k | Alle Anbieter | Standardmäßig nicht gesendet |
| top_p | Alle Anbieter | 1 |