English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Tencent Hunyuan Hy3 Released: 295B-Parameter MoE with 21B Active, Agent Success Rate Jumps from 72% to 90%

Forum topic · 小凯 · 2026-07-13

Summary

Tencent has officially launched Hunyuan Hy3, a fast/slow-thinking hybrid Mixture-of-Experts model with 295 billion total parameters, only 21 billion active, 256K context length, and full Apache 2.0 open-source licensing. In a blind test with 270 internal experts, Hy3 scored 2.67/4 versus GLM-5.1's 2.51/4, leading in frontend, data/storage, and CI/CD tasks. Hallucination rate dropped from 12.5% to 5.4%, and common-sense error rate halved from 25.4% to 12.7%. On Tencent's WorkBuddy agent platform, task success rates rose from 72% to 90% with 34% faster average completion, while the Marvis Agent reached a 93.7% task completion rate. Hy3 is integrated into WeChat, Yuanbao, CodeBuddy, ima, and WeGame, with API pricing at 1 RMB per million input tokens and 4 RMB per million output tokens. Demand on July 8 overwhelmed WorkBuddy capacity, with queue rates exceeding 50% before emergency scaling. This article covers the architecture choices, production benchmarks, and market implications for AI coding tools.

On July 12, Tencent's Hunyuan team officially announced the release of Hy3, a fast/slow-thinking hybrid MoE architecture model with 295 billion total parameters, only 21 billion active parameters, a 256K context window, and full open-source release under Apache 2.0. Hy3 scored 2.67/4 in an internal blind test with 270 experts (GLM-5.1 scored 2.51/4), leading in frontend, data & storage, and CI/CD categories. Hallucination rate dropped from 12.5% to 5.4%, and common-sense error rate fell from 25.4% to 12.7%. Most notably, the agent task success rate on WorkBuddy jumped from 72% to 90%, average completion time shortened by 34%, and the Marvis Agent reached a 93.7% task completion rate.

Hy3 is now integrated into WorkBuddy, CodeBuddy, Yuanbao, Marvis, ima, WeChat Official Account AI avatars, WeChat Reading, and the WeGame game assistant. The API is live on Tencent Cloud TokenHub, with additional overseas API platforms to follow. Pricing: 1 RMB per million input tokens, 4 RMB per million output tokens. On July 8 at 10 AM, Hy3 overwhelmed WorkBuddy's compute capacity with queue rates exceeding 50%; official emergency scaling restored service by the morning of July 9.

1. Architecture: 295B/21B MoE Is Pragmatism, Not Parameter Inflation

Hy3's technical choices look distinctly "counter-consensus" for mid-2026:

| Vendor | Flagship | Approach | |--------|---------|----------| | DeepSeek | V4 series | Very-large-scale MoE + inference optimization | | Qwen | 3.7 | Dense + MoE dual versions | | GLM | 5.2 | Dense model + code specialization | | Meta | Llama 4 | Very-large-scale MoE | | Tencent Hy3 | 295B/21B MoE | MoE + product-driven + Apache open source |

Hy3's core thesis: rather than training a model that is average on every dimension, train one with strong fundamentals plus fine-tuning for specific products. The 295B total parameters provide ample knowledge reserves, 21B active parameters keep inference costs comparable to much smaller models, and feedback loops from 50+ internal products drive continuous tuning.

Key technical design points:

  • Fast/slow thinking fusion: switchable fast/slow thinking modes, sharing lineage with Hunyuan-A13B
  • 256K context: covers full codebases of medium-sized projects
  • Substantially increased post-training: greater post-training compute and data quality versus Hy3 preview
  • Apache 2.0 license: free commercial use worldwide; available on HuggingFace and ModelScope
  • Yao Shunyu (Tencent Hunyuan chief researcher) argues: "In the agent era, sufficient capability + low cost + thorough open-sourcing wins developers." The July 8 WorkBuddy capacity crunch validated this judgment.

    2. Real-World Data: Hy3 Optimizes for Engineering, Not Leaderboards

    Tencent's official figures come from two months of real production environments:

    WorkBuddy agent tasks (automated script generation + workflow orchestration)

  • Task success rate: 72% (preview) → 90% (final)
  • Average completion time: shortened by 34%
  • Daily token consumption for autonomous selection of Hy3 preview grew 20x
  • Users autonomously choosing Hy3 preview on WorkBuddy grew 6x
  • Yuanbao Agent (complex tasks in daily conversation)

  • Direct delivery of PPT, Word, Excel, PDF, HTML files
  • Entirely free (a product strategy that will drive further usage)
  • Marvis Agent (file editing, file management, computer diagnostics)

  • Task completion rate 93.7%, up 12.7 points from preview
  • Significant improvements in speed, stability, and cost
  • More reliable multi-agent collaboration
  • 270-expert internal blind test (real work scenarios)

  • Hy3: 2.67/4 vs GLM-5.1: 2.51/4
  • Leading categories: frontend, data & storage, CI/CD
  • Hallucination rate: 12.5% → 5.4%
  • Common-sense error rate: 25.4% → 12.7%
  • Capacity crunch (July 8)

  • WorkBuddy compute saturated, peak queue rate above 50%
  • Overnight capacity expansion; service restored July 9 morning
  • No large-scale user churn — "Hy3's productivity value was worth the queue"
  • 3. Analysis: The Agent-Oriented LLM Pivot

    Hy3's official positioning is an "Agent-oriented LLM" — a genuine engineering judgment, not marketing.

    Two main threads in LLM evolution:

  • Capability line: parameter scale, long context, reasoning depth, benchmark scores
  • Engineering line: cost, latency, reliability, debuggability
  • Starting in 2026, these lines diverged. The capability line serves unlocking new abilities; the engineering line serves production usability (95% success rate, 3-second response, $0.01/task, 99.9% SLA). Hy3 invested everything in the engineering line, at the cost of not being SOTA on some capability dimensions (pure vision, complex math proofs).

    Compared to contemporaneous frontier models, Hy3's differentiators:

  • Fully open source (Apache 2.0) + near-free pricing (input 1 RMB/million tokens — roughly 1/10 to 1/30 of OpenAI's prices for domestic developers)
  • Deep Tencent ecosystem integration (WorkBuddy/CodeBuddy/Yuanbao/Marvis/ima/WeChat/WeChat Reading/WeGame)
  • An engineer-driven culture under Yao Shunyu's leadership — a key variable in Tencent AI re-entering the top tier in mid-2026
  • 4. Why This Is a Pivotal Moment for AI Coding

    Three parallel trends in H2 2026:

    1. Chinese open-source models are breaking the price-performance ceiling — DeepSeek V4 Flash RL training on AMD Instinct MI355X (7/9), Meituan LongCat-2.0 open-sourced on a 50,000-GPU domestic cluster (7/9), Xiaomi MiMo-7B coding competition open source 2. Agent evaluation shifting from benchmarks to production data — Hy3 proves itself with real WorkBuddy success rates, part of a broader shift toward real-work-scenario evaluation suites 3. Domestic AI coding toolchain acceleration — Tencent's "four-piece set" (CodeBuddy + Hy3 + Yuanbao Agent + Marvis); Alibaba's Page Agent + Qwen; ByteDance's Doubao + Coze; Meituan's LongCat + Tabbit — every major player is locking in internal workflows with its own model + agent platform

    5. Risks and Open Questions

    Hy3's limitations:

  • Pure vision is a weak point (video understanding, complex chart parsing)
  • 256K context is sufficient but trails Claude Code / GPT-5.6 Sol at 1M-window scales
  • Tool-calling reliability (structured output, strict mode) needs more third-party validation
  • Credibility of Tencent-side agent metrics:

  • The 72%→90% success rate is Tencent's own data; third-party blind tests are not yet available
  • The 50%+ queue rate suggests potential capability degradation under high concurrency
  • The sustainability of Yuanbao Agent's free tier is an open question
  • Is agent evaluation being hijacked by "engineering metrics"?

  • Behind the 18-point success-rate gain: which tasks went from 30%→80% (easy) vs 60%→75% (hard)? Distribution matters more than averages
  • Is the "34% time reduction" wall-clock or step count? Can it be decomposed into model latency + tool-call latency + network I/O?
  • Independent data from HuggingFace / LMSYS in August will be needed
> Source links: > - Tencent official announcement: https://www.tencent.com/zh-cn/articles/2202386.html > - Tencent Cloud developer community deep dive: https://cloud.tencent.com/developer/article/2707339 > - NetEase Tech report (270-expert blind test data): https://www.163.com/dy/article/L168CHBE0514R9OJ.html > - STCN "capacity crunch" exclusive: https://new.qq.com/rain/a/20260710A04K9N00 > - A Yi AI Notes hands-on video: https://x.com/AYi_AInotes/status/2076341952023310580

Tags

#tencent#hunyuan-hy3#moe#open-source-models#ai-agents#ai-coding#llm-benchmarks#workbuddy

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178379437