Gemma 3 27B Instruct
GoogleGemmaOpen WeightGemma · Uso Comercial
Descripción
Gemma 3 27B is a 27-billion-parameter vision-language model from Google, handling text and image input and generating text output. It features a 128K context window, multilingual support, and open weights. Suitable for complex question answering, summarization, reasoning, and image understanding tasks.
Fecha de lanzamiento
2025-03-12
Parámetros
27.0B
Longitud del contexto
262K
Modalidades
image, text
Radar de capacidades
25
general
12
coding
34
reasoning
28
science
27
agents
70
multimodal
Rankings
| Dominio | #Posición | Puntuación | Fuente |
|---|---|---|---|
| Ranking de codificación | 495 | 11.0 | AA |
| Ranking general | 464 | 26.0 | AA |
| Ranking multimodal | 54 | 44.0 | LS |
| Ciencia | 461 | 27.0 | AA |
Puntuaciones de benchmarks (LLM Stats)
(LLM Stats (zeroeval))Biology
GPQANYU + Cohere + Anthropic (2023)
42.4%Aut.
Code
HumanEvalOpenAI (2021)
87.8%Aut.
LiveCodeBench
29.7%Aut.
Factuality
FACTS Grounding
74.9%Aut.
SimpleQA
10.0%Aut.
Finance
MMLU-Pro
67.5%Aut.
General
IFEvalGoogle Research (2023)
90.4%Aut.
Natural2Code
84.5%Aut.
Global-MMLU-Lite
75.1%Aut.
MBPP
0.74 / 100Aut.
MMMU (val)
64.9%Aut.
BIG-Bench Extra Hard
19.3%Aut.
MRCR v2 (8-needle)
13.5%Aut.
Image To Text
DocVQADocVQA (2020)
86.6%Aut.
VQAv2 (val)
71.0%Aut.
TextVQA
65.1%Aut.
Language
BIG-Bench Hard
87.6%Aut.
WMT24++
53.4%Aut.
ECLeKTic
16.7%Aut.
Math
GSM8k
95.9%Aut.
MATH
89.0%Aut.
MathVista-Mini
67.6%Aut.
HiddenMath
60.3%Aut.
Multimodal
AI2D
84.5%Aut.
ChartQAMasry et al. (2022)
78.0%Aut.
InfoVQA
70.6%Aut.
Reasoning
Bird-SQL (dev)
54.4%Aut.
Índices de evaluación AA
(Artificial Analysis)Math Index(Artificial Analysis)20.7
Coding Index(Artificial Analysis)10.1
Intelligence Index(Artificial Analysis)7.4
Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))0.9
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))0.7
Gpqa(NYU + Cohere + Anthropic (2023))0.4
Ifbench(Google Research (2023))0.3
Aime(MAA (Mathematical Association of America))0.3
Scicode(UIUC + Argonne National Lab (2024))0.2
Aime 25(MAA (Mathematical Association of America))0.2
Livecodebench(UC Berkeley + MIT + Cornell (2024))0.1
Tau2(Sierra + U Toronto + Vector Institute (2025))0.1
Lcr(Artificial Analysis)0.1
Terminalbench V2 10.0
Hle(Center for AI Safety + Scale AI (2025))0.0
Terminalbench Hard(Stanford × Laude Institute (2026))0.0
Tau Banking0.0
Puntuaciones por categoría LLM Stats
(LLM Stats (zeroeval))Structured Output90
Instruction Following90
Math80
Legal70
Multimodal70
Image To Text70
Finance70
Grounding70
Healthcare70
Vision70
Reasoning60
Language60
General60
Code60
Physics40
Factuality40
Biology40
Chemistry40
Long Context10
Precios
Precio de entradaGratis
Precio de salidaGratis
Precio mixto (3:1)Gratis
Precio de lectura caché$0.04 / 1M tokens
Velocidad
Tokens/seg0.0
Retraso del primer token0.00s
Tiempo hasta la respuesta0.00s
Ranking de Precios por Proveedor
Ranking de Precios por Proveedor
6 proveedores
Más barato: OpenRouterMás caro: STACKIT
ProveedorEntradaSalida
1OpenRouterMás barato
$0.08
$0.45
2Kilo Gateway
$0.08
$0.16
3Nebius Token Factory
$0.1
$0.3
4LLM Gateway
$0.1
$0.3
5NovitaAI
$0.119
$0.2
6STACKIT
$0.53
$0.76
Comparar precios entre diferentes proveedores de API para este modelo.