Gemini 2.5 Flash
everyais/gemini-2-5-flash
Gemini 2.5 Flash (everyais/gemini-2-5-flash) availability, capabilities, context limits, and public reference pricing on everyais.
Method
The page starts with current public catalog metadata and reference costs; operational metrics are added only when measured samples are available.
Source and update
GET /models/catalogUpdated
- Model type
- Chat
- Released
- 2025-06-17
- Context window
- 1.0M
- Price unit
- Per 1M tokens
- Maximum output
- 65,536
- Model reference price (USD)
Standard (≤200K)
- Input / 1M tokens
- $0.15
- Output / 1M tokens
- $0.6
- Cache read / 1M tokens
- $0.037
Long context (>200K, full request)
- Input / 1M tokens
- $0.3
- Output / 1M tokens
- $2.5
- Cache read / 1M tokens
- $0.03
Capabilities
- ✓ Streaming
- ✓ Tool use
- ✓ Vision
- ✓ JSON
- ✓ Reasoning
- ✓ Sampling controls
- ✓ Structured outputs
Supported endpoints
- /v1/chat/completions
Benchmarks
Evaluation results published by external sources. Only mappings approved by a human administrator are shown.
| Benchmark | Score | Source |
|---|---|---|
| frontiermath | 4.8% | Epoch AI Benchmarking HubCC-BYSource model label: gemini-2.5-flash |
| terminal-bench | 17.1% | Epoch AI Benchmarking HubCC-BYSource model label: gemini-2.5-flash |
Scores are reproduced as published; everyais does not re-measure them. Scores using different units cannot be compared. See benchmark sources and licenses.
Usage and availability trend
Last 8 days · 2026-09-03 Input 1,071 · Output 140 tokens.
Daily availability (success rate %)
Insufficient sample — not enough requests to publish a success rate.
Code example
Use the OpenAI SDK by changing only base_url.
from openai import OpenAI
client = OpenAI(
api_key="everyais_...",
base_url="https://api.everyais.com/v1",
)
response = client.chat.completions.create(
model="everyais/gemini-2-5-flash",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)Other models in this family
Compare models side by side
Compare capabilities, context windows, and reference prices to find a model for your use case. Your actual billed rate can differ by account.