NEWMEN
← back to all models

Google: Gemini 2.5 Flash

google/gemini-2.5-flash

Googleclosed-weight1.0M context4 providersIntelligence 71.6

Cheapest provider

$0.30 / 1M

Google

Fastest provider (p95)

182 tok/s

Google AI Studio

Intelligence (composite)

71.6

MMLU-Pro · HumanEval · math · GPQA

Per-provider performance

Latency / throughput / uptime measured across providers over the last 30 minutes of live traffic. Atlas’s router weighs these per call (with the eval-gate signal) when picking a variant for Standard and Batch tiers.

ProviderQuantInput $/1MOutput $/1MLatency p50 / p95Throughput p50 / p95Uptime 30mSuccess
Googlefull· unknown$0.3000$2.50001323ms / 8154ms58 / 164 tok/s90.65%96.2%
Googlefull· unknown$0.3000$2.50001323ms / 8154ms58 / 164 tok/s97.54%96.2%
Googlefull· unknown$0.3000$2.50001323ms / 8154ms58 / 164 tok/s96.34%96.2%
Google AI Studiofull· unknown$0.3000$2.5000959ms / 14999ms55 / 182 tok/s99.65%99.7%

“—” means live telemetry hasn’t accumulated enough recent traffic for that endpoint. “undisclosed” means the provider serves the model but doesn’t expose the quantization label (typically running fp8 / int8 internally).

Intelligence breakdown

Composite score is a weighted average of public benchmarks (30% MMLU-Pro, 25% code pass@1, 25% math, 20% GPQA). Numbers come from model cards and the Artificial Analysis intelligence harness; missing components are renormalised over what’s present.

MMLU-Pro

broad reasoning

Code

62.3

pass@1 (HumanEval / LiveCodeBench)

AIME 2025

72.0

math accuracy

GPQA Diamond

82.8

hard reasoning

Source: Gemini 2.5 Flash Thinking (GPQA 82.8, LiveCodeBench 62.3, AIME25 72.0); DeepSeek-R1-0528 comparison table

How Atlas routes Google: Gemini 2.5 Flash

  • Realtime tier — direct passthrough at the upstream’s native precision. Best for hard latency / quality guarantees.
  • Standard tier — Atlas picks the cheapest provider variant whose quantization has stayed green on your operation’s eval gates. For Google: Gemini 2.5 Flash that’s currently Google at $0.30/1M.
  • Batch tier — async, biggest discount. Roadmapped to use provider batch APIs (OpenAI / Anthropic) where available and queued spot capacity for open-weight workloads.