Reflectionbeam-1.0-501b-a23b

Reflection: Beam 1.0 501B A23B API Pricing

Beam 1.0 is Reflection's open-weight reasoning model for coding and long-horizon agentic workflows. It uses a sparse mixture-of-experts architecture with 501B total and 23B active parameters per to...

Base Input Price$0.30/ 1M tokens

Base Output Price$1.00/ 1M tokens

Prompt Cache Read$0.040-86.67% discount

Batch API RateN/A24h async offload

Inference Hosts & Aggregator Providers

Compare rates and latency across direct APIs, OpenRouter, and hyperscale cloud providers

Live verified rates

Provider Name	Input / 1M	Output / 1M	Prompt Caching	Access Endpoint
Reflection Direct	$0.30	$1.00	Supported	Connect
OpenRouter	$0.30	$1.00	Supported	Connect

Head-to-Head Comparisons for Reflection: Beam 1.0 501B A23B

Reflection: Beam 1.0 501B A23B vs OpenAI: GPT-6.1 Sol$0.3/M vs $2/M

Reflection: Beam 1.0 501B A23B vs GPT-4o (Omni)$0.3/M vs $2.5/M

Reflection: Beam 1.0 501B A23B vs Claude 3.5 Sonnet$0.3/M vs $3/M

Reflection: Beam 1.0 501B A23B vs DeepSeek-V3 (deepseek-chat)$0.3/M vs $0.27/M