# Claude Haiku 4.5 vs Gemini 3.7 Flash

> Side-by-side benchmark comparison for Claude Haiku 4.5 and Gemini 3.7 Flash across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/claude-haiku-4-5-vs-gemini-3-7-flash
- Last updated: September 27, 2026

- Shared sourced benchmarks: 5
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.7

## Quick Verdict

Pick Gemini 3.7 Flash if you want the stronger benchmark profile. Claude Haiku 4.5 only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- Gemini 3.7 Flash leads overall 67.66 to 42.48.
- The clearest category separation is in coding, where the averages are 19.8 for Claude Haiku 4.5 and 60.3 for Gemini 3.7 Flash.
- The biggest single benchmark swing is LiveCodeBench (Vals) in Coding, with scores of 41.2% and 88.7%.
- Gemini 3.7 Flash is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- Gemini 3.7 Flash also has the larger context window at 1M.

## Model Snapshot

| Property | Claude Haiku 4.5 | Gemini 3.7 Flash |
|----------|----------|----------|
| Creator | Anthropic | Google |
| Type | Proprietary | Proprietary |
| Reasoning | Non-Reasoning | Reasoning |
| Context | 200K | 1M |
| Overall Score | 42.48 | 67.66 |
| Benchmarks Covered | 10 | 25 |
| Pricing (input/output) | $1.00 / $5.00 | $0.75 / $3.75 |

## Category Breakdown

### Agentic

- Winner: Gemini 3.7 Flash
- Claude Haiku 4.5 public-lane score: 21.7 (Supported · #85/105)
- Gemini 3.7 Flash public-lane score: 58.6 (Supported · #20/105)

| Benchmark | Claude Haiku 4.5 | Gemini 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| JobBench | 16.0% | Coming soon | Coming soon |
| Terminal-Bench 2.1 (Vals) | 43.8% | 77.5% | Gemini 3.7 Flash |
| Terminal-Bench 2.1 | Coming soon | 85.8% | Coming soon |
| Terminal-Bench 3.0 | Coming soon | 14.9% | Coming soon |
| AutomationBench | Coming soon | 30.4% | Coming soon |
| OSWorld 2.0 | Coming soon | 47.9% | Coming soon |
| Agents' Last Exam | Coming soon | 26.3% | Coming soon |
| ApprenticeBench | Coming soon | 16% | Coming soon |

### Coding

- Winner: Gemini 3.7 Flash
- Claude Haiku 4.5 public-lane score: 19.8 (Supported · #117/135)
- Gemini 3.7 Flash public-lane score: 60.3 (Supported · #17/135)

| Benchmark | Claude Haiku 4.5 | Gemini 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Verified | 73.3% | Coming soon | Coming soon |
| VulcanBench v3 | 76.2% | Coming soon | Coming soon |
| LiveCodeBench (Vals) | 41.2% | 88.7% | Gemini 3.7 Flash |
| SWE-bench (Vals) | 66.6% | 80.8% | Gemini 3.7 Flash |
| FrontierCode 1.1 Main | Coming soon | 43.6% | Coming soon |
| DeepSWE | Coming soon | 65.3% | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 85.8% | Coming soon |
| FrontierSWE v2 | Coming soon | 20.3% | Coming soon |

### Reasoning

- Winner: Not comparable
- Claude Haiku 4.5 public-lane score: Coming soon
- Gemini 3.7 Flash public-lane score: 77.8 (Unranked · 5 rankable rows)

| Benchmark | Claude Haiku 4.5 | Gemini 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| MRCR v2 64K-128K | Coming soon | 97% | Coming soon |
| ARC-AGI-1 | Coming soon | 95.50% | Coming soon |
| ARC-AGI-2 | Coming soon | 84.6% | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- Claude Haiku 4.5 public-lane score: Coming soon
- Gemini 3.7 Flash public-lane score: 82.6 (#10/50)

| Benchmark | Claude Haiku 4.5 | Gemini 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| CharXiv w/o tools | Coming soon | 84.5% | Coming soon |
| CharXiv | Coming soon | 88.7% | Coming soon |
| LVBench | Coming soon | 85.4% | Coming soon |

### Knowledge

- Winner: Directional only
- Claude Haiku 4.5 public-lane score: 34.5 (Estimated · #111/158)
- Gemini 3.7 Flash public-lane score: 71.1 (Supported · #10/158)

| Benchmark | Claude Haiku 4.5 | Gemini 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| GPQA Diamond (Vals) | 72.2% | 93.9% | Gemini 3.7 Flash |
| MMLU-Pro (Vals) | 78.7% | 90.1% | Gemini 3.7 Flash |
| HLE-Verified | Coming soon | 53.6% | Coming soon |
| LABBench2 | Coming soon | 82.1% | Coming soon |
| BioMysteryBench (human-solvable) | Coming soon | 87.1% | Coming soon |
| BioMysteryBench (human-difficult) | Coming soon | 43.5% | Coming soon |

### Mathematics

- Winner: Not comparable
- Claude Haiku 4.5 public-lane score: 28.8 (Unranked · 2 rankable rows)
- Gemini 3.7 Flash public-lane score: Coming soon

| Benchmark | Claude Haiku 4.5 | Gemini 3.7 Flash | Winner |
|-----------|-----------|-----------|--------|
| FrontierMath v2 (Tiers 1-3) | 5.903% | Coming soon | Coming soon |
| FrontierMath v2 (Tier 4) | 2.083% | Coming soon | Coming soon |

## FAQ

### Which is better overall, Claude Haiku 4.5 or Gemini 3.7 Flash?

Gemini 3.7 Flash is ahead overall on BenchLM right now.

### Where is the biggest gap between Claude Haiku 4.5 and Gemini 3.7 Flash?

The widest category gap is in coding, where the averages are 19.8 for Claude Haiku 4.5 and 60.3 for Gemini 3.7 Flash.

### How many benchmarks does Claude Haiku 4.5 cover on BenchLM?

Claude Haiku 4.5 currently has 10 sourced benchmark scores on BenchLM.

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

Gemini 3.7 Flash currently has 25 sourced benchmark scores on BenchLM.

## Related Comparisons

- [Claude Haiku 4.5 vs Claude Haiku 4.5 Thinking](/compare/claude-haiku-4-5-vs-claude-haiku-4-5-thinking)
- [Gemini 3.7 Flash vs Claude Haiku 4.5 Thinking](/compare/claude-haiku-4-5-thinking-vs-gemini-3-7-flash)
- [Claude Haiku 4.5 vs GPT-6 Astra](/compare/claude-haiku-4-5-vs-gpt-6-astra)
- [Gemini 3.7 Flash vs GPT-6 Astra](/compare/gemini-3-7-flash-vs-gpt-6-astra)

## Explore More

- [Claude Haiku 4.5 profile](/models/claude-haiku-4-5)
- [Gemini 3.7 Flash profile](/models/gemini-3-7-flash)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)