Gemini 2.5 Pro
GoogleGeminiProprietary
Описание
A highly capable AI model from Google, designed for the agentic era. Gemini 2.5 Pro performs well on common benchmarks with enhanced reasoning, multimodal capabilities (text, image, video, audio input), and a 1M token context window.
Дата выхода
2025-06-05
Параметры
—
Длина контекста
1.0M
Модальности
audio, image, pdf, text, video
Радар способностей
38
general
51
coding
89
reasoning
59
science
78
agents
80
multimodal
Рейтинги
| Домен | #Место | Оценка | Источник |
|---|---|---|---|
| Аудио | 69 | 41.0 | AA |
| Рейтинг кодинга | 253 | 57.0 | AA |
| Общий рейтинг | 234 | 50.0 | AA |
| Мультимодальный рейтинг | 31 | 60.0 | LS |
| Наука | 178 | 62.0 | AA |
Оценки бенчмарков (LLM Stats)
(LLM Stats (zeroeval))Factuality
SimpleQA
54.0%Сам.
General
Global-MMLU-Lite
89.2%Сам.
Aider-Polyglot
82.2%Сам.
Aider-Polyglot Edit
72.7%Сам.
Long Context
MRCR
93.0%Сам.
MRCR 1M (pointwise)
82.9%Сам.
MRCR v2 (8-needle)
16.4%Сам.
Math
AIME 2024
92.0%Сам.
AIME 2025
88.0%Сам.
Multimodal
Video-MME
84.8%Сам.
VideoMMMU
83.6%Сам.
MMMU
82.0%Сам.
Vibe-Eval
67.2%Сам.
Reasoning
FACTS Grounding
87.8%Сам.
GPQANYU + Cohere + Anthropic (2023)
86.4%Сам.
LiveCodeBench v5
75.6%Сам.
LiveCodeBench
69.0%Сам.
SWE-Bench Verified
67.2%Сам.
Humanity's Last Exam
21.6%Сам.
ARC-AGI v2
4.9%
Индексы оценки AA
(Artificial Analysis)Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))96.7
Aime(MAA (Mathematical Association of America))88.7
Math Index(Artificial Analysis)87.7
Aime 25(MAA (Mathematical Association of America))87.7
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))86.2
Gpqa(NYU + Cohere + Anthropic (2023))84.4
Livecodebench(UC Berkeley + MIT + Cornell (2024))80.1
Lcr(Artificial Analysis)69.0
Tau2(Sierra + U Toronto + Vector Institute (2025))54.1
Ifbench(Google Research (2023))48.7
Scicode(UIUC + Argonne National Lab (2024))46.3
Coding Index(Artificial Analysis)33.3
Terminalbench V2 128.5
Terminalbench Hard(Stanford × Laude Institute (2026))26.5
Hle(Center for AI Safety + Scale AI (2025))22.5
Intelligence Index(Artificial Analysis)16.1
Tau Banking9.7
Оценки категорий LLM Stats
(LLM Stats (zeroeval))Language90
Long Context90
Multimodal80
Physics80
Healthcare80
Biology80
Chemistry80
Reasoning70
General70
Code70
Math60
Frontend Development60
Factuality50
Vision50
Spatial Reasoning0
Цены
Цена ввода$1.25 / 1M токенов
Цена вывода$10 / 1M токенов
Смешанная цена (3:1)$3.438 / 1M токенов
Цена чтения кэша$0.125 / 1M токенов
Скорость
Токенов/сек0.0
Задержка первого токена0.00s
Время до первого ответа0.00s
Рейтинг цен провайдеров
Рейтинг цен провайдеров
27 провайдеров
Самый дешевый: GoogleСамый дорогой: Cortecs
ПровайдерВводВывод
1GoogleСамый дешевый
$0
$0.00001
2DeepInfra
$0
$0.00001
3Poe
$0.87
$7
4Jiekou.AI
$1.125
$9
5302.AI
$1.25
$10
6NanoGPT
$1.25
$10
7Abacus
$1.25
$10
8Perplexity Agent
$1.25
$10
9OpenRouter
$1.25
$10
10ZenMux
$1.25
$10
11Kilo Gateway
$1.25
$10
12SAP AI Core
$1.25
$10
13Helicone
$1.25
$10
14FrogBot
$1.25
$10
15AIHubMix
$1.25
$10
16Vercel AI Gateway
$1.25
$10
17DevPass (LLM Gateway)
$1.25
$10
18Vertex
$1.25
$10
19FastRouter
$1.25
$10
20Auriko
$1.25
$10
21NEAR AI Cloud
$1.25
$10
22OrcaRouter
$1.25
$10
23Merge Gateway
$1.25
$10
24Ofox
$1.25
$10
25Modelis
$1.25
$10
26Impossibl
$1.25
$10
27Cortecs
$1.495
$9.964
Сравнение цен разных API-провайдеров для этой модели.