# Grok 4.20 vs Step 3.7 Flash

> Side-by-side benchmark comparison for Grok 4.20 and Step 3.7 Flash across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/grok-4-20-beta-vs-step-3-7-flash
- Last updated: October 15, 2026

- Shared sourced benchmarks: 4
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

Grok 4.20 leads by point estimate. Conditional score ranges do not establish rank confidence. Choose using the category evidence and your own workload tests.

## Summary

- Grok 4.20 has the higher overall point estimate 57.62 to 43.7. Conditional score ranges do not establish rank confidence.
- The clearest category separation is in reasoning, where the averages are 44.9 for Grok 4.20 and 76.5 for Step 3.7 Flash.
- The biggest single benchmark swing is DeepSearchQA in Agentic, with scores of 62.8% and 92.8%.
- Step 3.7 Flash is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- Grok 4.20 also has the larger context window at 2M.

## Model Snapshot

| Property | Grok 4.20 | Step 3.7 Flash |
|----------|----------|----------|
| Creator | xAI | StepFun |
| Type | Proprietary | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 2M | 256K |
| Overall Score | 57.62 | 43.7 |
| Benchmarks Covered | 23 | 11 |
| Pricing (input/output) | $1.25 / $2.50 | $0.20 / $1.15 |

## Category Breakdown

### Agentic

- Winner: Directional only
- Grok 4.20 public-lane score: 25.2 (Supported · #101/126)
- Step 3.7 Flash public-lane score: 37.1 (Estimated · #74/126)

| Benchmark | Grok 4.20 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 47.1% | Coming soon | Coming soon |
| DeepSearchQA | 62.8% | 92.8% | Step 3.7 Flash |
| Gert Labs | 38.36% | 51.57% | Step 3.7 Flash |
| Terminal-Bench 2.1 (Vals) | 44.2% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 59.5% | Coming soon |
| BrowseComp | Coming soon | 75.8% | Coming soon |
| Toolathlon | Coming soon | 49.5% | Coming soon |
| Claw-Eval | Coming soon | 67.1% | Coming soon |
| HLE w/ tools | Coming soon | 47.2% | Coming soon |

### Coding

- Winner: Directional only
- Grok 4.20 public-lane score: 21.6 (Supported · #117/149)
- Step 3.7 Flash public-lane score: 30.2 (Estimated · #95/149)

| Benchmark | Grok 4.20 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| LiveCodeBench Pro | 74.2% | Coming soon | Coming soon |
| SWE-bench Verified | 76.7% | Coming soon | Coming soon |
| SWE-bench Pro | 51.8% | 56.3% | Step 3.7 Flash |
| Vibe Code Bench | 4.06% | Coming soon | Coming soon |
| LiveCodeBench (Vals) | 84.3% | Coming soon | Coming soon |
| SWE-bench (Vals) | 72.2% | Coming soon | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 59.5% | Coming soon |

### Reasoning

- Winner: Not comparable
- Grok 4.20 public-lane score: 44.9 (Unranked · 2 rankable rows)
- Step 3.7 Flash public-lane score: 76.5 (Unranked · 2 rankable rows)

| Benchmark | Grok 4.20 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| ARC-AGI-2 | 53.3% | Coming soon | Coming soon |
| ARC-AGI-3 | 0.1% | Coming soon | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- Grok 4.20 public-lane score: 53.4 (#47/54)
- Step 3.7 Flash public-lane score: 75.4 (Unranked · 3 rankable rows)

| Benchmark | Grok 4.20 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| MMMU-Pro | 75.2% | Coming soon | Coming soon |
| CharXiv | 60.9% | Coming soon | Coming soon |
| ERQA | 54.1% | Coming soon | Coming soon |
| SimpleVQA | 57.4% | 79.2% | Step 3.7 Flash |
| MedXpertQA (MM) | 65.8% | Coming soon | Coming soon |
| V* | Coming soon | 95.3% | Coming soon |

### Knowledge

- Winner: Directional only
- Grok 4.20 public-lane score: 52.9 (Supported · #68/182)
- Step 3.7 Flash public-lane score: 45 (Estimated · #96/182)

| Benchmark | Grok 4.20 | Step 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| GPQA-D | 88.5% | Coming soon | Coming soon |
| HLE w/o tools | 31.6% | Coming soon | Coming soon |
| HealthBench Hard | 20.3% | Coming soon | Coming soon |
| MedXpertQA (Text) | 50.2% | Coming soon | Coming soon |
| GPQA Diamond (Vals) | 88.6% | Coming soon | Coming soon |
| MMLU-Pro (Vals) | 86.3% | Coming soon | Coming soon |

### Instruction Following

- Winner: Not comparable
- Grok 4.20 public-lane score: Coming soon
- Step 3.7 Flash public-lane score: 82.1 (#54/146)

## FAQ

### Which is better overall, Grok 4.20 or Step 3.7 Flash?

Grok 4.20 has the higher overall point estimate. Conditional score ranges do not establish rank confidence.

### Where is the biggest gap between Grok 4.20 and Step 3.7 Flash?

The widest category gap is in reasoning, where the averages are 44.9 for Grok 4.20 and 76.5 for Step 3.7 Flash.

### How many benchmarks does Grok 4.20 cover on BenchLM?

Grok 4.20 currently has 23 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

- [Grok 4.20 vs Grok 4.20 Multi-agent](/compare/grok-4-20-beta-vs-grok-4-20-multi-agent-beta)
- [Step 3.7 Flash vs Grok 4.20 Multi-agent](/compare/grok-4-20-multi-agent-beta-vs-step-3-7-flash)
- [Grok 4.20 vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-grok-4-20-beta)
- [Step 3.7 Flash vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-step-3-7-flash)

## Explore More

- [Grok 4.20 profile](/models/grok-4-20-beta)
- [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)