Hy4 preview
by Tencenttext->text7 endpoints
Neura Intelligence Index
82.4/ 100
Rank
#36—
Confidence
mediumBenchmarks
5(13% coverage)
Overview
Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that require planning, context continuity, and sustained multi-step execution.
Capabilities
Text generation
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
Top-P sampling
Stop sequences
Deterministic seed
Log probabilities
Modalities
Input
Text
Output
Text
Technical Specifications
Context Window
1.0M tokens
Max Output
64K tokens
Tokenizer
Other
Uptime (24h)
10000.0%
Parameters
770000000000B
License
apache_2_0Open Weights
Country
🇨🇳CN
Release Date
August 28, 20262mo
Supported Parameters
frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_completion_tokensmax_tokensmin_pparallel_tool_callspresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
Benchmark Performance
coding
| Benchmark | Score | Source |
|---|---|---|
| SWE-Bench Pro | 65.7 | verified |
reasoning
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 92.3 | verified |
agent
| Benchmark | Score | Source |
|---|---|---|
| MCP Atlas | 83.7 | verified |
| Toolathlon | 74.1 | verified |
| APEX Agents | 37.1 | verified |
Performance Over Time
Pricing
Per 1M tokens
Input$0.83
Output$2.50
Cache read$0.04
Added
August 28, 2026
Last synced 11/7/2026