
Hy4 flash llega a $0.070 por millón de tokens de entrada: el modelo Hy más barato de Tencent es también el que menos documentación tiene
- GoogleNUEVOGoogle: Gemini 4 Carbon2026-11-10$2.00 / $10.00 por 1M de tokens · 90 tok/s
- AnthropicNUEVOAnthropic: Claude Fable 5.22026-11-09$8.00 / $40.00 por 1M de tokens · 97 tok/s
- orcaNUEVOOrca: OrcaC2 1.02026-11-0717 tok/s
- OrcaNUEVOOrcaCyber: GLM 5.3 Abliterated / Uncensored2026-11-04$3.00 / $5.00 por 1M de tokens · 87 tok/s
- OrcaNUEVOOrcaCyber: Hy4 Abliterated / Uncensored2026-11-04$0.80 / $1.60 por 1M de tokens · 133 tok/s
- OrcaNUEVOOrcaCyber: Kimi K3.1 Pro Abliterated / Uncensored2026-11-04$4.00 / $20.00 por 1M de tokens · 72 tok/s
- OrcaOrcaCyber: MiniMax M3.1 Flash Abliterated / Uncensored2026-10-29$0.80 / $1.60 por 1M de tokens · 130 tok/s
- OrcaOrcaCyber: GLM 5.3 Flash Abliterated / Uncensored2026-10-27$0.80 / $1.60 por 1M de tokens · 65 tok/s
- OrcaOrcaCyber: Qwen3.8 Flash Next Abliterated / Uncensored2026-10-26$0.80 / $1.60 por 1M de tokens · 92 tok/s
- OrcaOrcaCyber: MiMo V2.6 Flash Abliterated / Uncensored2026-10-26$0.80 / $1.60 por 1M de tokens · 43 tok/s
- MoonshotMoonshotAI: Kimi K3.1 Pro2026-10-2447Inteligencia
- OrcaOrcaCyber: DeepSeek V4.1 Flash Abliterated / Uncensored2026-10-23$0.80 / $1.60 por 1M de tokens · 193 tok/s
- minimaxMiniMax: M3.1 Flash2026-10-1929Inteligencia
- OrcaOrca: OrcaCyber Ultra 1.02026-10-17$5.00 / $10.00 por 1M de tokens · 150 tok/s
- deepseekDeepSeek: DeepSeek V4.1 Flash 10172026-10-1733Inteligencia
- DeepSeekDeepSeek: DeepSeek V4.1 Flash2026-10-1740Inteligencia
- OrcaOrca: OrcaCyber Zero 1.52026-10-10$3.00 / $7.50 por 1M de tokens · 119 tok/s
- openaiOpenAI: GPT-6.1 Sol2026-09-2955Inteligencia
- anthropicAnthropic: Claude Sonnet 5.52026-09-2855Inteligencia
- typesafeTypeSafe: Jev 1.132026-09-24$0.04 / $0.00 por 1M de tokens · 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.
Lo que realmente dice la tarjeta de tarifas
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 entrada / ¥2 salida / ¥0.02 en caché, 224B en total con 12B activos.
• Tencent HY4 — ¥6 input / ¥18 output / ¥0.3 cached, 770B total with 49B active, the general-availability build from 4 November 2026.
• Vista previa de Hy4: la misma tarifa de ¥6 / ¥18 / ¥0,3, Apache 2.0, pesos descargables desde el 28 de agosto de 2026.
Eso supone una reducción de doce veces en la entrada y de nueve veces en la salida frente a su propio buque insignia, de una generación a la siguiente, sin cambios en la ventana de contexto. La entrada en caché cae quince veces, que es la partida que determina la economía de cualquier agente que reenvíe un prompt de sistema largo en cada paso. Para una carga de trabajo que lee un documento grande una vez y luego le hace veinte preguntas, la tasa de caché es la tasa que importa, y 0,003 $ por millón tiene un precio que hace que ese patrón resulte casi gratis.
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.
De dónde vienen 12 mil millones de parámetros activos

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.
Detalle de arquitectura. Ninguno publicado. El número de capas, el número de expertos, la anchura de enrutamiento, el tamaño del codificador visual y si la multimodalidad es nativa o está añadida a posteriori: todo ello está ausente. Las cifras de 224B y 12B provienen de la página de investigación; todo lo que es más granular no existe en público.
La prueba a ciegas y qué abarca «coincide con 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.
Lee con atención la formulación, porque el post de lanzamiento y la página de investigación no dicen lo mismo. El post dice que el modelo "iguala" a Claude Haiku 5.5. La página dice que la satisfacción quedó cerca de la de este. Cerca no es igual, y un proveedor que resume su propio panel elegirá el enfoque que más le favorezca. Más importante aún: el panel está formado por 85 empleados de la empresa que creó el modelo, trabajando en tareas que la empresa seleccionó, en una interfaz que la empresa también desarrolla. Esa es una forma razonable de decidir si lanzar un modelo internamente. No es evidencia de que los resultados de un tercero vayan a coincidir, y no establece paridad en ningún tipo de tarea que las 27 categorías no incluyeran.
Lo que la prueba sí establece es más acotado y aun así útil: en tareas típicas de oficina —documentos, diapositivas, hojas de cálculo, ese tipo de cosas para las que sirve WorkBuddy—, un modelo de 12B activos no era obviamente peor que un competidor cercano a la frontera, a juicio de personas que usan ambos a diario. Si eso se generaliza, el argumento de coste es abrumador. Si se generaliza es la pregunta abierta, y no hay ningún benchmark público en el lanzamiento que pueda responderla.
Trazando el 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.
Lo que la versión no contiene
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.
No hay una tabla de benchmarks legible por máquina. Los resultados de DeepSWE, Terminal-Bench 4.0, APEX-Agents, AutomationBench, HLE y OSWorld 2.1 existen como imágenes de gráficos en la página del proveedor, no como texto. Es una elección extraña para un lanzamiento cuyo argumento central es la eficiencia medible, y significa que las cifras no se pueden copiar, comparar con un diff ni verificar sin leer píxeles.
Sin tarjeta de seguridad ni del sistema. Tampoco se incluye declaración alguna sobre el comportamiento de rechazo, el tratamiento de datos o la disponibilidad regional más allá de los productos de China continental mencionados.
Dónde se ejecuta y cómo probarlo de forma económica
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.
¿A qué apuntarlo primero?
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.

Eso encaja en un plan de pruebas sencillo. Toma veinte tareas que ya ejecutes, de las cuales la mayoría sean de documentos o de tipo agente y unas pocas sean verdaderamente difíciles de ingeniería. Registra los tokens de entrada, los tokens de salida, los aciertos de caché, el tiempo real y si la salida era utilizable. Ejecútalas en Hy4 flash y en la línea base que uses hoy. En una tarde sabrás si el 18x se aplica a tu tráfico, lo cual es más de lo que te dirá el propio lanzamiento.
El modelo Hy más barato es el menos legible.
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.
Comparados en este artículo3
Detectado en este artículo · Benchmarks: Artificial Analysis · actualizado a diario
