Hy4 preview Official Specifications, Pricing & Benchmarks
Hy4 preview (hy4-preview) by Tencent 腾讯混元, released 2026-08-28, featuring 1,048,576 context window, 64,000 max output tokens, and Text input.
📋 Hy4 preview Official API & System Card Specifications
Verified against official Tencent 腾讯混元 developer documentation & pricing
1. Identity & Capabilities
API Spec- API Model ID
- hy4-preview
- Release Date
- 2026-08-28
- Weight License
- Open-Weight (API Spec)
- Native Reasoning
- ✅ Supported
- Function Calling & JSON Schema
- ✅ 原生支持
- Batch API 异步批处理
- ✅ Supported (50% off)
2. Context, Output & Modalities
Limits- Max Context Window
- 1,049,000 tokens (1049K)
- Max Output Tokens
- 64,000 tokens
- Input Modalities
- Text
- Output Modalities
- Text
- Direct China Access
- ✅ Direct CN + RMB
- Deployment Mode
- Open-Weight + Cloud API
3. Official Pricing & Cache
USD / 1M- Input / 1M Tokens
- $ 0.834
- Cached Input / 1M
- $ 0.042 -95%
- Output / 1M Tokens
- $ 2.501
🎛️ Official Sampling & Reasoning Parameters
Recommended generation settings & API parameter constraints
📊 Hy4 preview Full Benchmark Scorecard
32 public benchmarks · ranks among 154 models · synced 2026-11-20
💻 Coding 11 items
Terminal-Bench 2.1 is an updated release of the Terminal-Bench benchmark that tests AI agents' ability to operate a computer via the terminal.
MCP Atlas is a benchmark for evaluating AI models on scaled tool use capabilities, measuring how well models can coordinate and utilize multiple tools across complex multi-step tasks.
A multilingual benchmark for issue resolving in software engineering that covers Java, TypeScript, JavaScript, Go, Rust, C, and C++.
CyberGym is a benchmark for evaluating AI agents on cybersecurity tasks, testing their ability to identify vulnerabilities, perform security analysis, and complete security-related challenges in a controlled environment.
SWE-Bench Pro is an advanced version of SWE-Bench that evaluates language models on complex, real-world software engineering tasks requiring extended reasoning and multi-step problem solving.
DeepSWE is a software engineering agent benchmark evaluated with the mini-swe-agent harness, where each task is solved in an isolated container with no internet access. It measures an agent's ability to autonomously resolve real-world coding issues end to end.
SkillsBench evaluates coding agents on self-contained programming tasks, measuring practical engineering skills across diverse software development scenarios.
NL2Repo evaluates long-horizon coding capabilities including repository-level understanding, where models must generate or modify code across entire repositories from natural language specifications.
PostTrainBench evaluates a model's ability to autonomously post-train base models. Given pretrain-only base models, the agent must complete the full pipeline of data synthesis, training, evaluation, and iteration within a time budget, scored across downstream benchmarks such as…
SWE-Marathon is an ultra-long-horizon software engineering benchmark covering tasks such as building compilers, optimizing kernels, and developing production-grade services. It measures whether agents can sustain quality across extremely long engineering trajectories.
Program Bench evaluates code-generation agents by asking them to recreate a program's behavior from only a compiled binary and documentation.
🤖 Agents & Tools 12 items
WideSearch is an agentic search benchmark that evaluates models' ability to perform broad, parallel search operations across multiple sources. It tests wide-coverage information retrieval and synthesis capabilities.
BankerToolBench is a public benchmark that evaluates models on banking and finance tool-use tasks. Models are scored against dataset rubrics, measuring their ability to correctly invoke tools and complete multi-step financial workflows.
DRACO is a deep research benchmark that evaluates an agent's ability to gather, synthesize, and reason over information to answer complex research questions. Scores are based on official rubrics per question, with the final score being the average across all questions.
Tool Decathlon is a comprehensive benchmark for evaluating AI agents' ability to use multiple tools across diverse task categories. It measures proficiency in tool selection, sequencing, and execution across ten different tool-use scenarios.
OfficeQA Pro evaluates AI models on professional knowledge-work questions and tasks drawn from real office workflows, including document analysis, spreadsheet reasoning, and information synthesis across business domains.
OneMillion Bench evaluates AI agents on high-economic-value tasks that require sustained, reliable execution across long-horizon real-world workflows.
Job Bench evaluates AI agents on realistic professional tasks that require multi-step planning, research, and production of work artifacts.
Workspace Bench evaluates AI agents on high-economic-value workplace tasks that span multi-step planning, file processing, and tool use across realistic office and productivity workflows.
GDPval-AA evaluates AI agents on economically valuable professional knowledge-work tasks and reports performance as an Elo score.
APEX-Agents is a benchmark evaluating AI agents on long horizon professional tasks that require sustained reasoning, planning, and execution across complex multi-step workflows.
AutomationBench is a tool-use benchmark that evaluates AI agents on automating real-world workflows, testing their ability to orchestrate tools and complete multi-step automation tasks.
Agents' Last Exam is a challenging benchmark for AI agents on hard, long-horizon tasks that test sustained reasoning, planning, and tool use, reported with and without tool access.
📐 Math 6 items
MathArena Apex is a challenging math contest benchmark featuring the most difficult mathematical problems designed to test advanced reasoning and problem-solving abilities of AI models. It focuses on olympiad-level mathematics and complex multi-step mathematical reasoning.
ArXivMath is a final-answer benchmark of research-level mathematics maintained by MathArena. Problems are extracted monthly from recent arXiv paper abstracts, then filtered through automated and manual checks to ensure they are self-contained, non-trivial, and verifiable.
Text-only Humanity's Last Exam variant evaluated with tool use enabled.
Text-only Humanity's Last Exam variant evaluated without tool use.
CritPT is a challenging reasoning benchmark reported by Qwen for evaluating frontier mathematical and critical problem-solving capability.
HorizonMath is an extremely difficult frontier mathematics benchmark designed to test the limits of mathematical reasoning on research-level and competition-beyond problems.
🧠 Reasoning & Science 3 items
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
BioMysteryBench evaluates a model's ability to reason through challenging molecular biology problems, reporting performance on a hard subset and on the subset of problems solved by human experts.
SuperChem is a benchmark of advanced chemistry problems requiring expert-level domain knowledge and reasoning.
Source: ZeroEval / llm-stats.com (✓ = third-party verified, others vendor-reported).
Closest Self-Hostable Open-Weight Alternatives to Hy4 preview
Open-weight models with full HuggingFace weights for private GPU / Mac deployment
Frequently Asked Questions
Is Hy4 preview open-source or proprietary?
Hy4 preview (hy4-preview) is a proprietary commercial model by Tencent 腾讯混元. Weights and internal layer dimensions are undisclosed; it is accessed via cloud API.
What are the context limit and API pricing of Hy4 preview?
Hy4 preview supports 1049K input context and 64,000 max output tokens, priced at $0.834/1M input ($0.042/1M cached) and $2.501/1M output.