Mesa de cartas modelo Gemini 3.6 Flash de DeepMind. · fuente
la mesa de cartas modelo Gemini 3.6 Flash de DeepMind; Grok 4.5 no reportado. · fuente
Mesa de cartas modelo Gemini 3.6 Flash de DeepMind. · fuente
| Modelo | Ingrese $/1 millón | Producción $/1 millón | SWE-Bench Pro (Pública) | Banco de terminales 2.1 | Banco MLE |
|---|---|---|---|---|---|
| Géminis 3.6 Flash | $1.50 | $7.50 | 58.7% | 78.0% | 63.9% |
| Géminis 3.5 Flash | $1.50 | $9.00 | 55.1% | 76.2% | 49.7% |
| Géminis 3.1 Pro | $2.00 | $12.00 | 54.2% | 73.8% | 42.6% |
| GPT-5.6 Luna | $1.00 | $6.00 | 62.7% | 84.7% | 47.6% |
| Grok 4.5 | $2.00 | $6.00 | 64.7% | 83.3% | 43.2% |
| Claudio Soneto 5 | $3.00 | $15.00 | 63.2% | 80.4% | 66.9% |
Como se indica en la tarjeta modelo de DeepMind; Los precios son sin almacenamiento en caché. · fuente
Estimate from the version's published specs, benchmarks and reviews, 0-5; scores are only comparable within one criteria year. Scored as launched (Gemini 3.6 Flash, 21 Jul 2026): Artificial Analysis Intelligence Index 50 (v4.1, high reasoning; same as 3.5 Flash; v4.1 was current at launch, AA's page now shows 34 on v4.3.2 so later re-indexing is not used); 1M-token input window (1,048,576 per Google API docs, 64K output); $1.50 in / $7.50 out per 1M = $3.00 blended at 3:1; 304 output tokens/s measured by Artificial Analysis pre-launch (its current page shows 202 t/s, a later snapshot); inputs text, image, video, audio/PDF. Coding null: AA's index at launch used Terminal-Bench v2.1 (Google reports 78.0% on v2.1), no Terminal-Bench 4.0 figure was published at launch, so the 4.0 anchor is not applied. Reviews are AA's own pre-launch benchmark note, OfficeChai, a comparison blog (emergent.sh) and 9to5Google quoting Google; no hands-on tests found, so 4 sources are mostly benchmark-led. blog.google
Gemini 3.6 Flash scores 50 on the Artificial Analysis Intelligence Index, matching Gemini 3.5 Flash
Gemini 3.6 Flash lands at exactly 50 points on version 4.1 of the index, the identical score its predecessor was already sitting at
Gemini 3.6 Flash is an efficiency upgrade, not an intelligence upgrade.
Gemini 3.6 Flash "delivers higher precision with fewer unwanted code edits and reduced execution loops"
| Gemini 3.6 Flash and Gemini 3.5 Flash-Lite: Halving Time per Task · 21 jul 2026 | 0.00 |
|---|---|
| Gemini 3.6 Flash Scores 50 On Artificial Analysis Intelligence Index, Same As Gemini 3.5 Flash · 21 jul 2026 | 0.00 |
| Google launches Gemini 3.6 Flash and 3.5 Flash-Lite, teases Gemini 4 · 21 jul 2026 | 0.40 |
| Gemini 3.6 Flash vs Gemini 3.5 Flash: What Changed and Should You Upgrade? · 24 jul 2026 | -0.10 |
Fotos de lectores, separadas de los medios con fuentes del pin. Valora lo que comiste o viste y, en un pin de restaurante, etiqueta los platos directamente en tu foto.
La primera entrada es siempre la fuente del pin. La confianza global es una media ponderada de la firmeza con la que cada referencia respalda las horas de inicio y fin usadas arriba; una referencia cuenta la mitad por cada 180 días que sea más antigua que la más reciente.
El reportaje de lanzamiento de 9to5Google está fechado el 21 de julio de 2026 y dice que Google «anunció hoy Gemini 3.6 Flash y 3.5 Flash-Lite»; la ficha del modelo 3.6 Flash de DeepMind[2] se publicó el 21 de julio de 2026.
DeepMind's model card, published 21 July 2026, lists the benchmark table and the $1.50 and $7.50 per million token prices.
MarkTechPost's same-day write-up of the three-model Flash tier release.
Neural Newscast's episode "Google Ships Three New Gemini Flash Models [Model Behavior]" (2026-07-21) says: "with enterprise software. [00:48] Thatcher Collins: I'm Thatcher Collins. [00:49] Thatcher Collins: Nina, the timing of these releases is certainly noteworthy. [00:53] Thatcher Collins: Today , Google shipped Gemini 3.6 Flash, 3.5 Flash Lite, and a specialized security-focused version called Flash Cyber. [01:03] Thatcher Collins: This blitz comes immediately after"[1]
¿Falta algo o hay un error? Dilo con tus palabras: un enlace que respalde este pin, otra fecha de inicio o fin y por qué, o un dato que falta o está mal. La IA lo contrasta con las fuentes del pin, busca otras mejores y añade cualquier página que te dé la razón. Las fuentes del propio pin siguen contando más. La IA también mira las imágenes: una que muestra otra cosa, o la muestra mal, pasa al final o se sustituye.
Título original: Gemini 3.6 Flash Released
| Benchmark | Gemini 3.6 Flash | Gemini 3.5 Flash | Gemini 3.1 Pro | GPT-5.6 Luna | Grok 4.5 | Claude Sonnet 5 |
|---|---|---|---|---|---|---|
| Precio de entrada ($/1M de tokens, sin caché) | $1.50 | $1.50 | $2.00 | $1.00 | $2.00 | $3.00 |
| Precio de salida ($/1M de tokens) | $7.50 | $9.00 | $12.00 | $6.00 | $6.00 | $15.00 |
| SWE-Bench Pro (Public) (tareas variadas de programación agéntica) | 58.7% | 55.1% | 54.2% | 62.7% | 64.7% | 63.2% |
| DeepSWE v1.1 (ingeniería de software de largo horizonte) | 49% | 37% | 12% | 67% | 54% | 54% |
| Terminal-bench 2.1 (programación agéntica en terminal, arnés Terminus-2) | 78.0% | 76.2% | 73.8% | 84.7% | 83.3% | 80.4% |
| MLE-Bench (ingeniería de aprendizaje automático) | 63.9% | 49.7% | 42.6% | 47.6% | 43.2% | 66.9% |
| GDPVal-AA v2 (trabajo del conocimiento, Elo) | 1421 | 1349 | 965 | 1584 | 1535 | 1607 |
| OSWorld-Verified (uso agéntico del ordenador) | 83.0% | 78.4% | 76.2% | 72.6% | — | 81.2% |
| CharXiv Reasoning (síntesis de información a partir de gráficos complejos, sin herramientas) | 85.2% | 84.2% | 83.3% | 82.7% | 81.6% | 77.0% |
| CharXiv Reasoning (síntesis de información a partir de gráficos complejos, con herramientas) | 89.4% | 84.9% | 83.2% | — | — | 88.3% |
| GDM-MRCR v2 (8-needle) (rendimiento en contexto largo, 128k (media)) | 91.8% | 77.3% | 84.9% | 74.8% | 81.4% | 71.6% |
| GDM-MRCR v2 (8-needle) (rendimiento en contexto largo, 1M (puntual)) | 54.0% | 26.6% | 26.3% | — | — | — |