LLM Reference

Gemini 2.5 Pro

Released
2025-06-17
Last refreshed
2026-06-29
Status
Researched 42d ago
ProprietaryCommercial use: conditionalMultimodalCodingRAGAgentsLong contextVisionClassificationJSON / Tool use

Gemini 2.5 Pro is worth evaluating for coding, rag, and agents when its provider route and context window match the workload.

Use it for

  • Teams evaluating coding, rag, and agents
  • Workloads that can use a 1m context window
  • Buyers comparing 4 tracked provider routes

Do not use it for

  • Workloads where another current model has stronger sourced task evidence
Specifications
Released
2025-06-17
Context
1m
Max output
65,536
Architecture
Decoder Only
Knowledge cutoff
2025-01
Specialization
general
Openness
Proprietary
License
ProprietaryCommercial use: conditional
Weights
Not released
Code
Unknown
Training
Pretrained
Created by

Pioneering artificial intelligence research.

London, United Kingdom
Founded 2014
Website
Pricing
Output / 1M
$10.00
Input / 1M
$1.25

Cheapest of 4 routes · GCP Vertex AI · cache read $0.125

About

Google DeepMind's most capable Gemini 2.5 model with native thinking/reasoning support. Features a 1M-token context window, multimodal inputs (text, image, audio, video), function calling, and strong performance across coding, mathematics, and scientific reasoning tasks.

Gemini 2.5 Pro is a proprietary model in the Gemini 2.5 family. The structured metadata tracks a 1m-token context window, multimodal input, reasoning, function calling, tool use, structured outputs, and code execution. This page tracks provider routes through Google AI Studio, GCP Vertex AI, OpenRouter, and 1 more, with the cheapest tracked route listed at $1.25 input and $10 output per 1M tokens. Headline tracked benchmarks include HumanEval 93.1, SWE-bench Verified 63.8, and LiveCodeBench 75.6.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ D

4 relevant benchmarks in the decision map.

RAG

Included by capability and metadata signals in the decision map.

Agents

Q/$ D

1 relevant benchmark in the decision map.

Provider price ladder

Compare all 4

Compare API pricing across 4 providers for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MBatch in / outCacheRoute
GCP Vertex AI$1.25$10.00$0.625 / $5.00read $0.125
Serverless
Google AI Studio$1.25$10.00$0.625 / $5.00read $0.125
Serverless
OpenRouter$1.25$10.00--
Serverless
Vercel AI Gateway$1.25$10.00-read $0.125
Serverless

Available via routers & gateways(13)

Capabilities

VisionMultimodalReasoningFunction CallingTool UseStructured OutputsCode ExecutionPrompt CachingBatch API

Benchmark peer barsfor Coding

Benchmark scores(12)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionSource
HumanEval93.12025-03https://storage.googleapis.com/deepmind-media/gemini/gemini_v2_5_report.pdf
SWE-bench Verified63.8custom agent scaffold, extended thinkinghttps://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/
LiveCodeBench75.6v5 pass@1, Google I/O 2025https://blog.google/technology/google-deepmind/google-gemini-updates-io-2025/
Aider Polyglot83.1preview-06-05 (32k think)https://aider.chat/docs/leaderboards
Chatbot Arena1398.0https://lmarena.ai
Google-Proof Q&A86.4GPQA Diamond, GA June 2025 model cardhttps://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-2-5-Pro-Model-Card.pdf
MMLU PRO86.2https://huggingface.co/spaces/TIGER-Lab/MMLU-Pro
AIME 202492.0https://deepmind.google/technologies/gemini/pro/
Massive Multi-discipline Multimodal Understanding81.7standard MMMU pass@1, no extended thinkinghttps://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/
AIME 202586.7AIME 2025 pass@1, extended thinkinghttps://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/
Terminal-Bench 2.032.6Terminal-Bench 2.0, Terminus 2 agenthttps://www.tbench.ai/leaderboard/terminal-bench/2.0
Humanity's Last Exam18.8Humanity's Last Exam, no tools, extended thinkinghttps://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/

Migration checks

No linked migration route is available for this model yet.

Compare Gemini 2.5 Pro with other models

Show all 63 popular comparisonssorted by 7-day search impressions
Gemini 2.5 Pro vs Claude Sonnet 4.6783Gemini 2.5 Pro vs Claude Opus 4.7596Gemini 2.5 Pro vs o3 Deep Research571Gemini 2.5 Pro vs Kimi K2.5450Gemini 2.5 Pro vs Qwen3.6-27B224Gemini 2.5 Pro vs GLM-5.1212Gemini 2.5 Pro vs Grok 4.3185Gemini 2.5 Pro vs GLM-5180Gemini 2.5 Pro vs GPT-5.5159Gemini 2.5 Pro vs Claude Haiku 4.5154Gemini 2.5 Pro vs GPT-5.4142Gemini 2.5 Pro vs DeepSeek V3.1138Gemini 2.5 Pro vs GPT-5.2129Gemini 2.5 Pro vs Grok-3112Gemini 2.5 Pro vs Qwen2.5-72B-Instruct111Gemini 2.5 Pro vs Claude Opus 4.6102Gemini 2.5 Pro vs Claude Opus 4.5102Gemini 2.5 Pro vs Gemma 7B Instruct74Gemini 2.5 Pro vs Claude Sonnet 4.556Gemini 2.5 Pro vs Llama 3 70B Instruct49Gemini 2.5 Pro vs DeepSeek V3 Base44Gemini 2.5 Pro vs Trinity-Large-Thinking43Gemini 2.5 Pro vs GPT-5.2 Codex43Gemini 2.5 Pro vs GPT-5.4-Cyber41Gemini 2.5 Pro vs Mistral Large 240Gemini 2.5 Pro vs Qwen3.5-35B-A3B40Gemini 2.5 Pro vs Phi-3 Mini 4k37Gemini 2.5 Pro vs Kimi K2 Instruct36Gemini 2.5 Pro vs Qwen3.6-35B-A3B32Gemini 2.5 Pro vs Llama 3.2 1B Instruct31Gemini 2.5 Pro vs DeepSeek V3.229Gemini 2.5 Pro vs o3 Mini27Gemini 2.5 Pro vs DeepSeek V326Gemini 2.5 Pro vs Code Davinci 00125Gemini 2.5 Pro vs Mixtral 8x7B23Gemini 2.5 Pro vs Phi 3.5 Mini Instruct20Gemini 2.5 Pro vs Llama 3.1 70B Instruct18Gemini 2.5 Pro vs Llama 3.1 405B Instruct17Gemini 2.5 Pro vs DeepSeek R1 052814Gemini 2.5 Pro vs Mistral Nemotron14Gemini 2.5 Pro vs Qwen3.5-27B11Gemini 2.5 Pro vs Qwen3-Max11Gemini 2.5 Pro vs Qwen3.5-122B-A10B10Gemini 2.5 Pro vs Phi-4 Mini Flash Reasoning10Gemini 2.5 Pro vs Trinity-Large-Preview8Gemini 2.5 Pro vs Qwen3-9B8Gemini 2.5 Pro vs Mistral Large 3 675B Instruct8Gemini 2.5 Pro vs GPT-4o-mini Search Preview8Gemini 2.5 Pro vs Llama Guard 3 1B6Gemini 2.5 Pro vs Mixtral 8x22B Instruct v0.36Gemini 2.5 Pro vs Llama 3 8B Instruct6Gemini 2.5 Pro vs Llama 2 70B Chat6Gemini 2.5 Pro vs Phi-4 Reasoning Vision 15B5Gemini 2.5 Pro vs Qwen2-7B-Instruct5Gemini 2.5 Pro vs ShieldGemma 9B5Gemini 2.5 Pro vs Llama 2 13B Chat5Gemini 2.5 Pro vs Qwen3-235B-A22B5Gemini 2.5 Pro vs Grok 2 Vision4Gemini 2.5 Pro vs Mistral Magistral Small 25093Gemini 2.5 Pro vs Qwen3.5-397B-A17B3Gemini 2.5 Pro vs GLM-5V-Turbo3Gemini 2.5 Pro vs Grok 3 Mini2Gemini 2.5 Pro vs Magistral Small 25061

Frequently asked questions

What is the context window of Gemini 2.5 Pro?

Gemini 2.5 Pro has a context window of 1m tokens.

What is the max output of Gemini 2.5 Pro?

Gemini 2.5 Pro can generate up to 65,536 output tokens.

How much does Gemini 2.5 Pro cost?

Gemini 2.5 Pro is available at $1.25/1M input tokens through Google AI Studio.

When was Gemini 2.5 Pro released?

Gemini 2.5 Pro was released on 2025-06-17.

Which providers offer Gemini 2.5 Pro?

Gemini 2.5 Pro is available from 4 providers: Google AI Studio, GCP Vertex AI, OpenRouter, Vercel AI Gateway.

What benchmarks has Gemini 2.5 Pro been tested on?

Gemini 2.5 Pro has been evaluated on 12 benchmarks, including HumanEval, SWE-bench Verified, LiveCodeBench, Aider Polyglot, Chatbot Arena.