# Beam 1.0 501B-A23B FP8 vs Step 3.7 Flash

> Side-by-side benchmark comparison for Beam 1.0 501B-A23B FP8 and Step 3.7 Flash across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/reflection-beam-fp8-vs-step-3-7-flash
- Last updated: October 16, 2026

- Shared sourced benchmarks: 6
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

Beam 1.0 501B-A23B FP8 leads by point estimate. Conditional score ranges do not establish rank confidence. Choose using the category evidence and your own workload tests.

## Summary

- Beam 1.0 501B-A23B FP8 has the higher overall point estimate 71 to 43.64. Conditional score ranges do not establish rank confidence.
- The clearest category separation is in reasoning, where the averages are 85.8 for Beam 1.0 501B-A23B FP8 and 76.5 for Step 3.7 Flash.
- The biggest single benchmark swing is Terminal-Bench 2.1 in Agentic, with scores of 78.9% and 59.5%.
- Beam 1.0 501B-A23B FP8 is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- Beam 1.0 501B-A23B FP8 also has the larger context window at 1M.

## Model Snapshot

| Property | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash |
|----------|----------|----------|
| Creator | Reflection AI | StepFun |
| Type | Open Weight | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 1M | 256K |
| Overall Score | null | 43.64 |
| Benchmarks Covered | 19 | 11 |
| Pricing (input/output) | Pricing unavailable / Pricing unavailable | $0.20 / $1.15 |

## Category Breakdown

### Agentic

- Winner: Not comparable
- Beam 1.0 501B-A23B FP8 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 36.8 (Estimated · #76/128)

| Benchmark | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| τ³-bench results | 36.8% | Coming soon | Coming soon |
| HLE w/ tools | 51.1% | 47.2% | Beam 1.0 501B-A23B FP8 |
| Terminal-Bench 2.1 | 78.9% | 59.5% | Beam 1.0 501B-A23B FP8 |
| MCP Atlas | 81.5% | Coming soon | Coming soon |
| BrowseComp | 76.7% | 75.8% | Beam 1.0 501B-A23B FP8 |
| DeepSearchQA | 79.6% | 92.8% | Step 3.7 Flash |
| Toolathlon | Coming soon | 49.5% | Coming soon |
| Claw-Eval | Coming soon | 67.1% | Coming soon |
| Gert Labs | Coming soon | 51.57% | Coming soon |

### Coding

- Winner: Not comparable
- Beam 1.0 501B-A23B FP8 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 30.2 (Estimated · #96/150)

| Benchmark | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| DeepSWE | 43.0% | Coming soon | Coming soon |
| SWE-bench Pro | 65.9% | 56.3% | Beam 1.0 501B-A23B FP8 |
| Terminal-Bench 2.1 | 78.9% | 59.5% | Beam 1.0 501B-A23B FP8 |
| SWE-Atlas Codebase QnA | 32.2% | Coming soon | Coming soon |
| SWE Multilingual | 78.1% | Coming soon | Coming soon |
| SWE-bench Verified | 80.2% | Coming soon | Coming soon |
| SciCode | 48.3% | Coming soon | Coming soon |

### Reasoning

- Winner: Not comparable
- Beam 1.0 501B-A23B FP8 public-lane score: 85.8 (Unranked · 3 rankable rows)
- Step 3.7 Flash public-lane score: 76.5 (Unranked · 2 rankable rows)

| Benchmark | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| CritPt | 17.4% | Coming soon | Coming soon |
| LongBench v2 | 67.6% | Coming soon | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- Beam 1.0 501B-A23B FP8 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 75.4 (Unranked · 3 rankable rows)

| Benchmark | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| SimpleVQA | Coming soon | 79.2% | Coming soon |
| V* | Coming soon | 95.3% | Coming soon |

### Knowledge

- Winner: Not comparable
- Beam 1.0 501B-A23B FP8 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 44.9 (Estimated · #97/184)

| Benchmark | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| GPQA-D | 90.3% | Coming soon | Coming soon |
| HLE w/o tools | 36% | Coming soon | Coming soon |

### Instruction Following

- Winner: Directional only
- Beam 1.0 501B-A23B FP8 public-lane score: 89.8 (#28/148)
- Step 3.7 Flash public-lane score: 82.1 (#56/148)

| Benchmark | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| IFBench | 80.7% | Coming soon | Coming soon |

### Mathematics

- Winner: Not comparable
- Beam 1.0 501B-A23B FP8 public-lane score: 87.9 (Unranked · 1 rankable row)
- Step 3.7 Flash public-lane score: Coming soon

| Benchmark | Beam 1.0 501B-A23B FP8 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| AIME26 | 98.0% | Coming soon | Coming soon |

## FAQ

### Which is better overall, Beam 1.0 501B-A23B FP8 or Step 3.7 Flash?

Beam 1.0 501B-A23B FP8 has the higher overall point estimate. Conditional score ranges do not establish rank confidence.

### Where is the biggest gap between Beam 1.0 501B-A23B FP8 and Step 3.7 Flash?

The widest category gap is in reasoning, where the averages are 85.8 for Beam 1.0 501B-A23B FP8 and 76.5 for Step 3.7 Flash.

### How many benchmarks does Beam 1.0 501B-A23B FP8 cover on BenchLM?

Beam 1.0 501B-A23B FP8 currently has 19 sourced benchmark scores on BenchLM.

### How many benchmarks does Step 3.7 Flash cover on BenchLM?

Step 3.7 Flash currently has 11 sourced benchmark scores on BenchLM.

## Related Comparisons

- [Beam 1.0 501B-A23B FP8 vs Beam 1.0 501B-A23B](/compare/reflection-beam-vs-reflection-beam-fp8)
- [Step 3.7 Flash vs Beam 1.0 501B-A23B](/compare/reflection-beam-vs-step-3-7-flash)
- [Beam 1.0 501B-A23B FP8 vs Beam 1.0 501B-A23B NVFP4](/compare/reflection-beam-fp8-vs-reflection-beam-nvfp4)
- [Step 3.7 Flash vs Beam 1.0 501B-A23B NVFP4](/compare/reflection-beam-nvfp4-vs-step-3-7-flash)

## Explore More

- [Beam 1.0 501B-A23B FP8 profile](/models/reflection-beam-fp8)
- [Step 3.7 Flash profile](/models/step-3-7-flash)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)