
Hy4 flash arriva a 0,070 $ per milione di token in input: il modello Hy più economico di Tencent è anche quello meno documentato
- GoogleNUOVOGoogle: Gemini 4 Carbon2026-11-10$2.00 / $10.00 per 1M di token · 90 tok/s
- AnthropicNUOVOAnthropic: Claude Fable 5.22026-11-09$8.00 / $40.00 per 1M di token · 97 tok/s
- orcaNUOVOOrca: OrcaC2 1.02026-11-0717 tok/s
- OrcaNUOVOOrcaCyber: GLM 5.3 Abliterated / Uncensored2026-11-04$3.00 / $5.00 per 1M di token · 87 tok/s
- OrcaNUOVOOrcaCyber: Hy4 Abliterated / Uncensored2026-11-04$0.80 / $1.60 per 1M di token · 133 tok/s
- OrcaNUOVOOrcaCyber: Kimi K3.1 Pro Abliterated / Uncensored2026-11-04$4.00 / $20.00 per 1M di token · 72 tok/s
- OrcaOrcaCyber: MiniMax M3.1 Flash Abliterated / Uncensored2026-10-29$0.80 / $1.60 per 1M di token · 130 tok/s
- OrcaOrcaCyber: GLM 5.3 Flash Abliterated / Uncensored2026-10-27$0.80 / $1.60 per 1M di token · 65 tok/s
- OrcaOrcaCyber: Qwen3.8 Flash Next Abliterated / Uncensored2026-10-26$0.80 / $1.60 per 1M di token · 92 tok/s
- OrcaOrcaCyber: MiMo V2.6 Flash Abliterated / Uncensored2026-10-26$0.80 / $1.60 per 1M di token · 43 tok/s
- MoonshotMoonshotAI: Kimi K3.1 Pro2026-10-2447Intelligenza
- OrcaOrcaCyber: DeepSeek V4.1 Flash Abliterated / Uncensored2026-10-23$0.80 / $1.60 per 1M di token · 193 tok/s
- minimaxMiniMax: M3.1 Flash2026-10-1929Intelligenza
- OrcaOrca: OrcaCyber Ultra 1.02026-10-17$5.00 / $10.00 per 1M di token · 150 tok/s
- deepseekDeepSeek: DeepSeek V4.1 Flash 10172026-10-1733Intelligenza
- DeepSeekDeepSeek: DeepSeek V4.1 Flash2026-10-1740Intelligenza
- OrcaOrca: OrcaCyber Zero 1.52026-10-10$3.00 / $7.50 per 1M di token · 119 tok/s
- openaiOpenAI: GPT-6.1 Sol2026-09-2955Intelligenza
- anthropicAnthropic: Claude Sonnet 5.52026-09-2855Intelligenza
- typesafeTypeSafe: Jev 1.132026-09-24$0.04 / $0.00 per 1M di token · 219 tok/s
Tencent released Hy4 flash on 16 November 2026, and the specification reads like a cost-engineering exercise more than a capability one: 224 billion total parameters of which 64 billion are Engram parameters, 12 billion active per token, native multimodality, a 1,048,576-token context, and a mainland list price of ¥0.5 per million input tokens against ¥2 per million output. The company's framing is that it matches Claude Haiku 5.5 on everyday work in a blind test run on its own employees, at roughly an eighteenth of the per-task cost. The blind test is real and the price is real; the comparison across them is Tencent's own arithmetic, and the difference between those two statements is most of what anyone deploying this model needs to know.
There is no model card. There is no architecture paper, no licence, no weights repository and no independent benchmark run. What Tencent published instead is a research page with benchmark results baked into PNG charts, a set of internal product testimonials, and a promise to open-source the model later. That is a thinner dossier than Hy4 preview shipped with in August, and the gap matters more here than it would for a flagship, because the entire pitch for Hy4 flash is efficiency — and efficiency claims are the ones you cannot check from a price table.
Cosa dice effettivamente il listino prezzi
Per million tokens, on the mainland card: ¥0.5 input, ¥2 output, ¥0.02 cached input. Tencent quotes the same model in dollars as $0.070 input, $0.281 output and $0.003 cached, which is the figure the launch post led with. Set it beside the rest of the family and the positioning is unambiguous.
• Hy4 flash — ¥0,5 input / ¥2 output / ¥0,02 in cache, 224B totali con 12B attivi.
• Tencent HY4 — ¥6 input / ¥18 output / ¥0.3 cached, 770B total with 49B active, the general-availability build from 4 November 2026.
Anteprima di Hy4 — la stessa tariffa di ¥6 / ¥18 / ¥0.3, Apache 2.0, pesi scaricabili dal 28 agosto 2026.
Si tratta di una riduzione di dodici volte sull'input e di nove volte sull'output rispetto alla propria ammiraglia, da una generazione alla successiva, senza alcuna modifica alla finestra di contesto. L'input in cache cala di quindici volte, ed è la voce che decide l'economia di qualsiasi agente che reinvia un lungo prompt di sistema a ogni passaggio. Per un carico di lavoro che legge un grande documento una volta e poi gli rivolge venti domande, la tariffa della cache è quella che conta, e $0,003 per milione è prezzato per rendere quel pattern quasi gratuito.
Two promotions sit on top of the list price and both are temporary. Tencent Cloud's TokenHub is running a 50% discount for a limited period with no stated end date, and WorkBuddy is free to use until 30 November. Neither should be modelled into a production budget; the list rate is the number that survives December.
Da dove provengono 12 miliardi di parametri attivi

The parameter split is the first thing to understand, because 64 billion Engram parameters inside a 224-billion-parameter model is not a normal arrangement. Engram, in Tencent's usage, is a memory-style component that holds learned associations rather than participating in every forward pass the way an expert does. The practical consequence is that a large share of the model's total size is not charged against per-token compute: 12 billion parameters activate per token, which is a quarter of what GLM-5.3 Flash activates and roughly a quarter of Tencent HY4's 49 billion.
Tencent names four levers, and each one attacks a different part of the bill.
Pre-training. Hybrid attention combined with LatentMoE and the Engram parameters. Tencent's stated result is a KV cache one twenty-fourth the size of Hy3's and one sixth of Hy4's. On long-context serving that is the difference between a request fitting on a node and a request needing a cluster, and it is the single most consequential number in the release for anyone renting GPUs by the hour.
Post-training. Self-distillation from Tencent HY4, then reinforcement learning. The stated outcome is a 50% reduction in end-to-end output length. This is where a flash model usually gives back part of its price advantage — shorter answers can mean shallower work — and it is also where the per-task cost claims below get most of their leverage. Output tokens are the expensive ones at $0.281 per million, so halving them halves the larger side of the invoice before the price cut is even applied.
Inference. A megakernel, four-bit quantization and a decoding path Tencent calls DFly, together sustaining better than 150 tokens per second of decode. Fast decode changes agent behaviour as much as it changes user experience: a tool-calling loop that spends less wall-clock per step can afford more steps inside the same latency budget.
Dettaglio sull'architettura. Nessun dettaglio pubblicato. Il numero di layer, il numero di esperti, l'ampiezza del routing, la dimensione dell'encoder visivo e se la multimodalità sia nativa o aggiunta a posteriori: tutto assente. Le cifre 224B e 12B provengono dalla pagina di ricerca; tutto ciò che è più dettagliato non esiste in pubblico.
Il test in cieco, e cosa comprende "pari a Claude Haiku 5.5"
The claim that travelled is the blind test. Tencent ran it with 85 users of its WorkBuddy product across 27 task categories and 259 real daily-work tasks, comparing Hy4 flash with several unnamed alternatives and reporting that user satisfaction came out close to Claude Haiku 5.5, that Hy4 flash had the lowest per-task cost of any model in the panel, and that it finished tasks in the shortest time.
Leggi bene la formulazione, perché il post di lancio e la pagina di ricerca non dicono la stessa cosa. Il post afferma che il modello "eguaglia" Claude Haiku 5.5. La pagina dice che la soddisfazione era vicina a quella. Vicino non significa uguale, e un fornitore che riassume il proprio panel sceglierà l'inquadratura che lo lusinga. Ancora più importante, il panel è composto da 85 dipendenti dell'azienda che ha creato il modello, che lavorano su compiti selezionati dall'azienda, in un'interfaccia costruita anch'essa dall'azienda. È un modo ragionevole per decidere se distribuire un modello internamente. Non è una prova che i risultati di terzi saranno gli stessi, e non stabilisce la parità su alcun tipo di compito che le 27 categorie non includevano.
Ciò che il test stabilisce è più circoscritto e comunque utile: sul lavoro di tipo d'ufficio — documenti, slide, fogli di calcolo, il genere di cose per cui è pensato WorkBuddy — un modello da 12B attivi non era palesemente peggiore di un concorrente prossimo alla frontiera, a giudizio di persone che li usano entrambi ogni giorno. Se questo è generalizzabile, la convenienza economica è schiacciante. Se lo sia è la questione aperta, e nella release non c'è alcun benchmark pubblico che possa rispondere.
Tracciamento del 18x
The "18x task cost reduction" in the launch post is not a price ratio. Hy4 flash costs twelve times less per input token than Tencent HY4, so a simple per-token comparison gives 12x, not 18x. The extra comes from Tencent's WorkIndex, an internal eight-suite benchmark covering PowerPoint creation, PowerPoint revision, human-machine co-writing, Markdown co-writing, document editing, table repair, audio and video processing, and visual understanding. On that suite, at the maximum tier and at busy-hour prices, Tencent reports per-task cost at one-thirteenth of Hy4 preview, one-twentieth of Tencent HY4, and one-eighth of DeepSeek-V4.1-Flash-1017.
One-twentieth against its own flagship and one-eighteenth against the model the post names is the same order of magnitude, so the 18x is a defensible round number — but it is a per-task figure, which means it folds in output length, retry rate and step count alongside price. Two of those three inputs are things Tencent controls through post-training and reports only as a percentage. A model that answers in half the tokens at a twelfth of the price is not 18x cheaper because of the price; it is 18x cheaper because it is also being made to talk less. If your task needs the longer answer, you get less than 18x, and the article you are reading cannot tell you how much less.
The same caveat applies to every per-task figure in the release, including the internal product numbers Tencent volunteered: a 70% cut in task time and roughly one-fifteenth the cost in WorkBuddy against Tencent HY4, a 2.52% general-chat gain against GLM-5.3-Flash in 元宝, 29% fewer output tokens and 6% fewer steps in 微信支付 agent flows against DeepSeek-V4.1-Flash-1017, 88% faster output and 60% fewer thinking tokens in 企业微信, and 22% to 39% fewer requests with 40% to 61% fewer tokens in 王者荣耀 root-cause analysis. These are the most credible numbers in the package, because they come from production systems with real users and real bills. They are also unaudited and selectively released, and no product publishes the flows where the model lost.
Cosa non contiene la release
No open weights. Tencent says it plans to open-source Hy4 flash and to iterate on it frequently. A plan is not a licence. Hy4 preview shipped under Apache 2.0 on day one with a Hugging Face repository, vLLM and SGLang recipes and published sampling defaults; Hy4 flash has none of that, and the wording around the open-source commitment is softer than the preview's was.
No independent evaluation. As of publication, Artificial Analysis has not posted a score for Hy4 flash, and no third party has reproduced a single number from the release. Every figure above is Tencent's.
Nessuna tabella di benchmark in formato leggibile da macchina. I risultati di DeepSWE, Terminal-Bench 4.0, APEX-Agents, AutomationBench, HLE e OSWorld 2.1 esistono come immagini di grafici sulla pagina del fornitore, non come testo. È una scelta strana per un lancio il cui intero argomento è l'efficienza misurabile, e significa che i numeri non possono essere copiati, sottoposti a diff o verificati senza leggere i pixel.
Nessuna scheda di sicurezza o del sistema. Inoltre, nessuna dichiarazione sul comportamento di rifiuto, sulla gestione dei dati o sulla disponibilità regionale oltre ai prodotti della Cina continentale citati.
Dove viene eseguito e come provarlo spendendo poco
Availability on launch day is Tencent's own surface: TokenHub on Tencent Cloud for API access, WorkBuddy for office tasks free until 30 November, and 元宝, 企业微信, QQ, ima and Marvis on the consumer and enterprise side. Nothing else is announced.

On OrcaRouter, Hy4 flash is not in the catalogue yet — the route that exists for this family is tencent/hy4-preview, at $0.834 per million input and $2.501 per million output, alongside a rate-limited free tier. That is the model Hy4 flash was distilled from, so it is the right baseline to measure against while you wait: run the same tasks on both once the flash route lands and you will have the per-task ratio that Tencent's WorkIndex implies without having to trust it.
The reason to care about where it is served is the price model. OrcaRouter passes provider list price through at 0% markup, which is not a discount but a commitment about what happens when a vendor cuts one: Tencent has already run two promotional discounts on this model in its first week, and a pass-through route reflects the next one the day it appears rather than when a reseller updates a table. Failover matters for the same reason in reverse — an unproven 12B-active model on launch-week infrastructure is exactly the model you want to be able to fall back from without rewriting a client, and the same key reaches more than 200 models if it does.
Su cosa puntarlo per prima cosa
Tencent's own deployment guidance splits the family cleanly, which is more useful than most vendor advice because it concedes a weakness: Hy4 flash for deep search, document processing, PowerPoint generation, simple frontend work and computer use; Tencent HY4 for long-horizon software engineering. The concession is in the last clause — the cheap model is not being sold as a coding replacement for the expensive one, and the internal WorkBuddy data shows a small drop on long coding tasks even as it reaches parity on daily work.

Ciò si traduce in un semplice piano di test. Prendi venti attività che già esegui, la maggior parte delle quali di tipo documentale o agentico e alcune davvero difficili dal punto di vista ingegneristico. Registra token di input, token di output, cache hit, tempo effettivo e se l'output era utilizzabile. Eseguile su Hy4 flash e sul baseline che usi oggi. Scoprirai in un pomeriggio se il 18x si applica al tuo traffico, il che è più di quanto ti dirà il rilascio stesso.
Il modello Hy meno costoso è quello meno leggibile.
Tencent has built something worth paying attention to: a quarter-active model with a million-token window, a KV cache small enough to change the serving maths, and a price that undercuts its own flagship by an order of magnitude. The engineering described — Engram parameters, a 50% output-length cut through self-distillation, four-bit inference at over 150 tokens per second — is the kind of work that shows up in a bill rather than a leaderboard, which is presumably why Tencent led with per-task cost.
But the release asks to be believed on exactly the quantities it declines to document. No weights, no model card, no independent score, benchmark numbers behind pixels, and a headline comparison drawn from a panel of its own employees. None of that is dishonest. All of it means the honest summary is "this is probably the best value in the Hy family and nobody outside Tencent has verified it," and the right response is a two-week trial with token counts, not a migration.
The open-source promise is the thing to diary. If Tencent ships weights the way it did for Hy4 preview — Apache 2.0, day one, with serving recipes — the efficiency claims become checkable and the model becomes a serious option for anyone who wants to run cheap inference themselves. Until that happens, Hy4 flash is a very good price attached to a very thin dossier, and the difference between those two facts is where your evaluation budget should go.
Confrontati in questo articolo3
Rilevato da questo articolo · Benchmark: Artificial Analysis · aggiornato ogni giorno
