NEWMEN
← back to all models

OpenAI: o3

openai/o3

OpenAIclosed-weight200K context1 providerIntelligence 79.3

Cheapest provider

$2.00 / 1M

OpenAI

Fastest provider (p95)

No throughput data yet — populated as traffic accumulates

Intelligence (composite)

79.3

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
OpenAIfull· unknown$2.0000$8.0000100.00%

“—” 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

85.9

broad reasoning

Code

58.6

pass@1 (HumanEval / LiveCodeBench)

AIME 2025

88.9

math accuracy

GPQA Diamond

83.3

hard reasoning

Source: OpenAI o3 (MMLU-Pro 85.9, GPQA 83.3, LiveCodeBench v6 58.6, AIME25 88.9); Qwen3-235B-A22B-Thinking-2507 comparison table

How Atlas routes OpenAI: o3

  • 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 OpenAI: o3 that’s currently OpenAI at $2.00/1M.
  • Batch tier — async, biggest discount. Roadmapped to use provider batch APIs (OpenAI / Anthropic) where available and queued spot capacity for open-weight workloads.