Beam 1.0 501B A23B
NEWReflection · Open weights · Released Oct 16, 2026 · reflection/beam-1.0-501b-a23b
Best $0.25 in / $1.00 out via Nebius · fp8
1.05M contextToolsReasoningStructured outputPrompt cachingOpen weights
Available from 4 providersUSD / MTOK · PRICES 19 MINUTES AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | UPTIME |
|---|---|---|---|---|---|---|---|---|
 Nebius | $0.25 | $1.00 | — | 1.05M | fp8 | — | — | 83.7% |
 Together | $0.40 | $1.20 | $0.04 | 1.05M | — | — | — | 89.5% |
 CoreWeave | $1.00 | $3.00 | $0.10 | 1.05M | fp4 | — | — | 87.1% |
 BaseTen | $1.30 | $4.30 | $0.13 | 1.05M | fp4 | — | — | 100.0% |
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Recent changesRSS ↗
Oct 17, 2026model added
About
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 token, and a...
Specifications
Context window1,048,576
Max output943,718
Modalities intext
Modalities outtext
LicenseOpen weights
ReleasedOct 16, 2026
Knowledge cutoff2026-06-30
TokenizerOther
WeightsHugging Face ↗
Open weights
safetensors
Embed badge
[](https://modelindex.ai/models/reflection/beam-1.0-501b-a23b)Sources
OpenRouter API19 minutes ago
AGGREGATORHugging Face19 minutes ago
COMMUNITYVerified 19 minutes ago · list prices, not negotiated rates.