Microsoft lists 2 models and Reflection 1. Microsoft is cheaper on a typical model ($0.35 per 1M tokens median). Below: line-up, price tiers, the cheapest model for each capability, and what each costs for five real workloads.
| Microsoft | Reflection | |
|---|---|---|
| Models listed | 2 | 1 |
| Models with published pricing | 2 | 1 |
| Typical price (median, 3:1 blend per 1M) | $0.35 | $1.50 |
| Cheapest paid model | Microsoft: Phi 4 ($0.09) | Reflection: Beam 1.0 501B A23B ($1.50) |
| Longest context | WizardLM-2 8x22B (66K) | Reflection: Beam 1.0 501B A23B (1.0M) |
| Free models | 0 | 0 |
| Newest model | Microsoft: Phi 4 (2025-01-10) | Reflection: Beam 1.0 501B A23B (2026-10-16) |
| Price changes in the last 90 days | 1 | 0 |
For each workload: the cheapest model from each provider that can actually do it (fits the context window; tool calling required for the agent), and its monthly cost at list price.
| Workload (per month) | Microsoft | Reflection |
|---|---|---|
| Customer-support chatbot 50,000 conversations — each a 6-turn conversation | $8.05 with Microsoft: Phi 4 | $135 with Reflection: Beam 1.0 501B A23B |
| RAG search over documents 20,000 questions — each a question answered from ~8 retrieved passages | $9.80 with Microsoft: Phi 4 | $150 with Reflection: Beam 1.0 501B A23B |
| Coding / tool-using agent 2,000 tasks — each a multi-step task with tool calls and file context | no suitable model | $156 with Reflection: Beam 1.0 501B A23B |
| Batch summarisation 10,000 documents — each a 6,000-word report summarised to a page | $6.44 with Microsoft: Phi 4 | $98.00 with Reflection: Beam 1.0 501B A23B |
| Long-document analysis 500 documents — each a 150-page contract or codebase read in one pass | no suitable model | $53.00 with Reflection: Beam 1.0 501B A23B |
| You need | Microsoft | Reflection |
|---|---|---|
| Any model | Microsoft: Phi 4 ($0.09) | Reflection: Beam 1.0 501B A23B ($1.50) |
| With tool calling | none listed | Reflection: Beam 1.0 501B A23B ($1.50) |
| With image input | none listed | none listed |
| With reasoning mode | none listed | Reflection: Beam 1.0 501B A23B ($1.50) |
| 128K+ context | none listed | Reflection: Beam 1.0 501B A23B ($1.50) |
| 1M+ context | none listed | Reflection: Beam 1.0 501B A23B ($1.50) |
| Model type | Microsoft | Reflection |
|---|---|---|
| text->text | 2 | 1 |
| Price per 1M (3:1 blend) | Microsoft | Reflection |
|---|---|---|
| Free | 0 | 0 |
| Under $0.50 | 1 | 0 |
| $0.50 – $3 | 1 | 1 |
| $3 – $10 | 0 | 0 |
| $10 and up | 0 | 0 |
Microsoft is cheaper on a typical model: a median of $0.35 per 1M tokens (3:1 input/output) against $1.50 for Reflection. The cheapest paid models are Microsoft: Phi 4 at $0.09 and Reflection: Beam 1.0 501B A23B at $1.50.
It depends on the job: Microsoft offers the broader line-up (2 models vs 1), Microsoft the lower typical price, Reflection the longest context (1.0M tokens). The workload table above shows which wins for chat, RAG, agents, batch jobs and long documents.
Microsoft: Microsoft lists 2 models and Reflection lists 1.
Reflection: Microsoft's longest is WizardLM-2 8x22B at 66K tokens; Reflection's is Reflection: Beam 1.0 501B A23B at 1.0M.
Neither lists a free model; both are pay-per-token.
Microsoft: 0 of 2 models accept images. Reflection: 0 of 1.