explainx.ai350K+ learn from us
TrendingAI News TodayPathwaysSkills
Pricing
explainx.ai

Upskill in AI — 16 free pathways, live workshops & bootcamps, and 50+ courses from practitioners. Plus the skills, tools, and MCP servers to practice on.

follow us

follow on google

Add explainx.ai as a preferred source

corporate training

support@explainx.ai

get started

Find your pathTake Free Evaluation

community

Join the community

learn

mind: share how you thinkpathways — start freeworkshopsbootcampscoursescompare Explainxcertificationsmock testsexplainx universitycorporate traininglearn skills & mcp

discover

skillsmcp serversexplainx mcpestacktoolsmdx readeragentsllmsdesignsdictionarypeopleagi trackerfelony benchranks

company

aboutvisionmissionteaminstructorsteach on explainxpartnershipscommunityhackathonscareers

content

daily AI newsstate of AI — live resultsblogreleasespromptsgeneratorsresource libraryfor LLMsexplainx.ai kids

solutions

all solutionsdeveloper upskillingmarketing upskillingproduct manager upskillingleadership upskilling

newsletter · weekly

Get AI news, tools, and insights in your inbox.

supportcontactprivacytermsdata rightshow we create contentsubmission guidelines

© 2026 AISOLO Technologies Pvt Ltd

explainx.ai

On this page

  • What is Xing4 in one sentence?
  • Ascend 910C at 8,192 cards: what the paper claims
  • Architecture knobs the authors stress
  • MindSpore HyperParallel: why the OS story matters
  • TeleGym: agent post-training as infrastructure
  • Weights, licensing, and what you can run today
  • How Xing4 fits export-control geography
  • Ascend 910C vs H100: what searchers should expect
  • Who should read this paper?
  • TeleGym vs Western agent gyms: a checklist
  • Long-sequence memory on Ascend clusters
  • Sparse activation and speculative decoding together
  • What to watch after the paper drop
  • CPU affinity and why host noise still matters at 8k GPUs
  • TeleChat lineage and open-weight expectations
  • MindSpore vs PyTorch for agent researchers
  • Related reading
← Back to blog

explainx / blog

Xing4 and TeleGym: Frontier Agentic Training on 8,192 Ascend 910C GPUs

Open Models, AI Hardware, Agents, Research, China AI

Part of AI Chips and Infrastructure

Xingchen AGI Lab's Xing4 report (arXiv 2611.22115) trains a 285B model on 8,192 Huawei Ascend 910C chips and ships TeleGym for agentic post-training on MindSpore.

Nov 18, 2026·8 min read·Yash Thakker
add explainx.ai
go deep
Xing4 and TeleGym: Frontier Agentic Training on 8,192 Ascend 910C GPUs

Xing4 is the newest technical report in the TeleChat line from Xingchen AGI Lab, published as arXiv 2611.22115 on November 17, 2026 and trending atop Hugging Face daily papers on November 18. It is not a consumer chat launch. It is a systems plus agent-training story: how a 285B-parameter run was staged on 8,192 Ascend 910C accelerators and how TeleGym wires environment rollouts into policy optimization.

What is Xing4 in one sentence?

Xing4 extends TeleChat with architecture search, MindSpore HyperParallel pretraining, and agentic post-training aimed at reasoning and tool use on Huawei Ascend — China's parallel stack to Nvidia CUDA clusters blocked or throttled by export rules.

Ascend 910C at 8,192 cards: what the paper claims

The abstract — mirrored on Hugging Face papers — lists concrete scale claims builders can verify only inside the PDF tables:

table · 3 cols
LayerClaim in abstractBuilder read
Hardware8,192 Ascend 910C cardsCluster-scale pretrain, not laptop fine-tune
Model285B parametersFrontier dense/sparse mix per arch section
FrameworkMindSpore HyperParallelTopology-aware hybrid sharding
Host tuningCPU affinity and core isolationReduces scheduler jitter on long jobs
OptimizerDistributed optimizer executionStandard at this scale; details in paper

explainx.ai does not have independent tokens-per-day measurements for this run. Treat throughput numbers as author-reported until third parties reproduce on Ascend racks. For earlier 910B co-design narratives, see DeepSeek and Huawei chip posts.

Daily briefFrom Yash · 350K+ students taught

Catch up on AI

One AI story explained in plain English, plus one thing to try. Every day, free. Unsubscribe anytime.

Architecture knobs the authors stress

Xing4 285B MoE architecture with MLA, DSA, and speculative decodingXing4 285B MoE architecture with MLA, DSA, and speculative decoding Image: Xingchen AGI Lab

Frontier labs now tune four coupled levers named in the abstract:

  1. Sparse activation — more capacity without paying full FLOPs every token.
  2. Attention design — memory bandwidth is the bottleneck on long agent trajectories.
  3. Depth vs width — trades routing complexity against parallel efficiency on Ascend meshes.
  4. Speculative decoding — inference-side speed that training stacks must not break.

That list mirrors what Western posts document for sparse attention co-design and speculative decoding guides, but here the target silicon is Ascend, not Blackwell.

MindSpore HyperParallel: why the OS story matters

The paper anchors pretraining on MindSpore HyperParallel with:

  • Topology-aware hybrid sharding
  • Coordinated long-sequence workloads
  • Activation memory management

If you only read model cards, you miss why Chinese frontier teams publish 138-author systems reports: export controls push full-stack ownership — chip, compiler, parallel framework, and agent gym — because they cannot assume NCCL on H100 will be available next quarter.

That is the same strategic line as Tencent HY4 on domestic supply chains, but Xing4 goes deeper on pretraining mechanics than a product blog.

TeleGym: agent post-training as infrastructure

TeleGym platform architecture for agent rollouts and policy optimizationTeleGym platform architecture for agent rollouts and policy optimization Image: Xingchen AGI Lab

TeleGym is the second half of the report. The abstract describes a unified agent execution platform that connects:

  • Execution in environments
  • Evaluation hooks
  • Trajectory construction
  • Policy optimization

Operational details that matter for agent builders:

  • Disaggregated asynchronous execution — rollouts do not block on slow tools uniformly.
  • Bounded oversampling — controls cost when tasks have heavy tails.
  • Partial rollout — keeps long-horizon trajectories usable across policy updates.

Partial rollout across policy updates in TeleGymPartial rollout across policy updates in TeleGym Image: Xingchen AGI Lab

Western readers should compare TeleGym's shape — not its branding — to SWE-CoWork's colleague simulator, Together compiler agents, and AWS phased agentic roadmap. Everyone is building harnesses that survive heterogeneous step times.

Weights, licensing, and what you can run today

On November 18, 2026, the Hugging Face paper page shows zero models and zero datasets citing 2611.22115. There is no public weight drop in the same bundle as the PDF.

Practical takeaway: Read Xing4 for architecture and training recipes; do not plan a Friday fine-tune until a model card lands with license terms. When weights appear, check whether inference requires Ascend-only kernels or ships CPU/GPU ports.

How Xing4 fits export-control geography

U.S. rules pushed advanced Nvidia sales out of direct China channels; Southeast Asian neocloud routing became a compliance battleground covered in Megaspeed reporting and congressional curb pushes.

Xing4 is the onshore response: train frontier-scale models where Ascend 910C supply is domestic. Whether that closes the quality gap with GPT-6.2 or Claude Opus 5.5 is an eval question, not settled in the abstract.

Ascend 910C vs H100: what searchers should expect

Search demand clusters on 910C vs H100 and 910C price. Public English sources in this report do not ship a reproducible MLPerf-style apples-to-apples table in the abstract.

Use three rules when you compare:

  1. Split training vs inference — Xing4 emphasizes pretrain plus agent post-train, not serving QPS.
  2. Split chip vs framework — MindSpore HyperParallel numbers do not automatically transfer to vLLM on H100.
  3. Split claimed vs verified — treat peak FLOPs as vendor math until an independent lab reruns the same config.

Who should read this paper?

table · 2 cols
AudienceWhy bother
Agent platform engineersTeleGym's async rollout design is portable as a checklist
Training infra leadsSharding plus host isolation patterns apply beyond Ascend
Policy readersDocuments China's alternate frontier stack under export pressure
Model shoppersWait for weights; abstract alone does not change your API picker

TeleGym vs Western agent gyms: a checklist

Use this table when you map TeleGym to your internal platform:

table · 3 cols
Design choiceTeleGym (per abstract)Typical Western RLHF stack
Rollout executorDisaggregated async workersOften synchronous env step loops
Task duration spreadBounded oversamplingFixed episode caps
Policy refreshPartial rollouts across updatesFull episode discard on version bump
Hardware targetAscend plus MindSporeCUDA plus PyTorch or JAX

None of these rows say which side wins on sample efficiency. They tell platform engineers what to ask vendors when buying an agent training cloud: do you preserve trajectory fragments across policy versions, and how do you bill long-tail tool calls?

Long-sequence memory on Ascend clusters

The abstract highlights coordinated management of long-sequence workloads and activation memory. Agent post-training pushes context lengths beyond chat pretraining defaults — tool transcripts, game boards, and code trees all inflate activations.

If you operate hybrid clouds, assume Ascend-optimized checkpoints will not drop into vLLM on H100 without a conversion project. Budget engineering quarters, not weekend quant jobs, when a lab publishes Ascend-only training recipes.

Sparse activation and speculative decoding together

Sparse MoE layers reduce active parameters per token; speculative decoding reduces latency per user request. Training stacks must co-design both or risk deployment surprises: a model cheap to train on paper may still thrash memory when agents spawn parallel rollouts.

Read Xing4's architecture section alongside NVIDIA sparse attention research and our KV cache explainer when you size inference for agent harnesses.

What to watch after the paper drop

Signals that Xing4 moved from report to product:

  • A Hugging Face model org under Xingchen-AGI with license text
  • TeleGym or MindSpore agent env repos with reproducible launch scripts
  • Third-party tokens per day claims on Ascend 910C at 285B scale
  • Government or telecom deployment posts referencing TeleChat branding

Until then, cite arXiv 2611.22115 only — not aggregator summaries.

CPU affinity and why host noise still matters at 8k GPUs

Large Ascend jobs remain host-bound for orchestration, dataloader fan-out, and logging. The abstract's CPU affinity and core isolation line is easy to dismiss as datacenter trivia. It is not. Scheduler jitter on the head node can desynchronize collective steps enough that effective MFU diverges from slides even when chips are nominal.

Western labs learned the same lesson on Slurm clusters running NCCL attribution failures. Xing4 is evidence Chinese frontier teams document host tuning beside model diagrams — a sign the race is systems-complete, not weights-only.

TeleChat lineage and open-weight expectations

TeleChat branding signals state-linked telecom research continuity. Readers hunting open weights should compare releases to Tencent HY4 plans and K2 Horizon fleets: some reports ship checkpoints, others stay paper-first.

When weights land, read licenses before fine-tune: Ascend-only inference clauses have appeared in other domestic model cards. Plan compliance before you port weights into a US-hosted vLLM cluster.

MindSpore vs PyTorch for agent researchers

You do not need Ascend hardware to extract design patterns from Xing4. Agent teams on CUDA can still copy TeleGym's async partial-rollout story when their env step latencies vary by orders of magnitude — a common pain in browser agent harnesses and MCP-heavy workflows. The chip stack differs; the scheduling problem does not.

Related reading

  • DeepSeek Huawei Ascend co-design claims
  • Tencent HY4 open-weights plan
  • China model distillation Reuters probe
  • Nvidia sparse attention co-design
  • Speculative decoding explainer
  • SWE-CoWork agent organizational benchmark

The author list spans well over a hundred names — typical for national-lab scale systems reports. That scale signals institutional commitment, not a weekend fine-tune drop.

If you benchmark Ascend clusters in procurement, ask vendors for reproducible job configs matching the paper's HyperParallel topology notes. Slides without config hashes rarely survive peer review in your own internal bake-offs.

Daily paper rank #1 on Hugging Face (November 18) reflects researcher attention, not a weight release. Treat rank as a literature signal, same as citation velocity on arXiv — not as proof the model beats your current API default.

Reading gets you the idea. Building it with feedback gets you the skill.

Spotted something out of date? Let us know.
Daily briefFrom Yash · 350K+ students taught

Catch up on AI

One AI story explained in plain English, plus one thing to try. Every day, free. Unsubscribe anytime.

  • ✓ Free
  • ✓ One email a day, no spam
  • ✓ Unsubscribe anytime
Yash Thakker

Written by

Yash Thakker

Yash is an AI expert with over 300K learners. Join his workshops →

View Yash Thakker in People in AI →

Related posts

Nov 11, 2026

ByteDance Frontier AI by 2030: Zhang Yiming's Timeline, Claimed vs Verified

On November 11, 2026, Bloomberg and relay outlets report that ByteDance co-founder Zhang Yiming told an internal AI meeting a personal agent like Meta Muse could ship in weeks, but matching OpenAI and Anthropic at the frontier may wait until 2029 or 2030 because Seed pretraining data lags rivals.

Nov 1, 2026

How Pro-China Are Chinese AI Models? Vals AI Tested 25 and Found It Is Complicated

A Vals AI study asked 25 Chinese, American and other models hundreds of politically sensitive questions about China. Results vary widely by model, language, system prompt, answer format and agent harness. Here is what it found and how to use it.

Nov 18, 2026

Can LLM Agents Invent Their Own Self-Improvement Loop? MIT Gameplay Study

Researchers asked whether frontier models still need humans to design self-improvement algorithms. Given only tools and a budget, stronger teachers write games, run A/B tests, and edit skill files that transfer to held-out benchmarks.