# GPT-5.4 vs LFM3-1.7B

> Side-by-side benchmark comparison for GPT-5.4 and LFM3-1.7B across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/gpt-5-4-vs-lfm3-1-7b
- Last updated: October 27, 2026

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

## Quick Verdict

GPT-5.4 leads by point estimate. Conditional score ranges do not establish rank confidence. Choose using the category evidence and your own workload tests.

## Summary

- GPT-5.4 has the higher overall point estimate 66.8 to 22.99. Conditional score ranges do not establish rank confidence.
- The clearest category separation is in knowledge, where the averages are 68.6 for GPT-5.4 and 25.5 for LFM3-1.7B.
- The biggest single benchmark swing is LiveCodeBench Pro in Coding, with scores of 87.5% and 31.7%.
- LFM3-1.7B is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- GPT-5.4 also has the larger context window at 1.05M.

## Model Snapshot

| Property | GPT-5.4 | LFM3-1.7B |
|----------|----------|----------|
| Creator | OpenAI | LiquidAI |
| Type | Proprietary | Open Weight |
| Reasoning | Reasoning | Reasoning |
| Context | 1.05M | 256K |
| Overall Score | 66.8 | 22.99 |
| Benchmarks Covered | 43 | 24 |
| Pricing (input/output) | $2.50 / $15.00 | $0.00 / $0.00 |

## Category Breakdown

### Agentic

- Winner: Directional only
- GPT-5.4 public-lane score: 54.3 (Supported · #46/134)
- LFM3-1.7B public-lane score: 14.4 (Estimated · #123/134)

| Benchmark | GPT-5.4 | LFM3-1.7B | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 75.1% | Coming soon | Coming soon |
| CyberGym | 79.0% | Coming soon | Coming soon |
| BrowseComp | 82.7% | Coming soon | Coming soon |
| OSWorld-Verified | 75% | Coming soon | Coming soon |
| MCP Atlas | 70.6% | 32.6% | GPT-5.4 |
| Toolathlon | 54.6% | Coming soon | Coming soon |
| τ²-bench results | 98.9% | Coming soon | Coming soon |
| Claw-Eval | 60.3% | 68.6% | LFM3-1.7B |
| DeepSearchQA | 73.6% | Coming soon | Coming soon |
| Gert Labs | 64.89% | Coming soon | Coming soon |
| ResearchClawBench | 15.3% | Coming soon | Coming soon |
| JobBench | 38.9% | Coming soon | Coming soon |
| ExploitGym | 6.0% | Coming soon | Coming soon |
| ApprenticeBench | 11% | Coming soon | Coming soon |
| MazeBench | 0% | Coming soon | Coming soon |
| AutomationBench | Coming soon | 41.3% | Coming soon |
| BFCL v4 | Coming soon | 71.4% | Coming soon |
| τ³-bench results | Coming soon | 68.0% | Coming soon |
| WebArena-Verified | Coming soon | 38.4% | Coming soon |
| ToolSandbox | Coming soon | 80.36% | Coming soon |
| HLE w/ tools | Coming soon | 13.9% | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 12.9% | Coming soon |
| LiveClaw | Coming soon | 45.60% | Coming soon |

### Coding

- Winner: Directional only
- GPT-5.4 public-lane score: 48 (Supported · #49/155)
- LFM3-1.7B public-lane score: 13 (Estimated · #142/155)

| Benchmark | GPT-5.4 | LFM3-1.7B | Winner |
|-----------|-----------|-----------|--------|
| LiveCodeBench Pro | 87.5% | 31.7% | GPT-5.4 |
| SWE-bench Pro | 57.7% | Coming soon | Coming soon |
| React Native Evals | 85.3% | Coming soon | Coming soon |
| Terminal-Bench 2.0 | 75.1% | Coming soon | Coming soon |
| Vibe Code Bench | 67.42% | Coming soon | Coming soon |
| PostTrainBench v1.1 | 19.0% | Coming soon | Coming soon |
| SciCode | Coming soon | 28.5% | Coming soon |
| SWE-bench Verified | Coming soon | 43.4% | Coming soon |
| SWE Multilingual | Coming soon | 23% | Coming soon |
| Terminal-Bench 2.1 | Coming soon | 12.9% | Coming soon |
| LiveCodeBench v6 | Coming soon | 73.5% | Coming soon |

### Reasoning

- Winner: Not comparable
- GPT-5.4 public-lane score: 64.2 (#19/26)
- LFM3-1.7B public-lane score: 61.3 (Unranked · 1 rankable row)

| Benchmark | GPT-5.4 | LFM3-1.7B | Winner |
|-----------|-----------|-----------|--------|
| ARC-AGI-2 | 74.0% | Coming soon | Coming soon |
| ARC-AGI-3 | 0.2% | Coming soon | Coming soon |
| ARC-AGI-1 | 93.67% | Coming soon | Coming soon |
| RULER (128K) | Coming soon | 65.56% | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- GPT-5.4 public-lane score: 71.5 (#35/58)
- LFM3-1.7B public-lane score: Coming soon

| Benchmark | GPT-5.4 | LFM3-1.7B | Winner |
|-----------|-----------|-----------|--------|
| ParseBench | 62.23% | Coming soon | Coming soon |
| MMMU-Pro | 81.2% | Coming soon | Coming soon |
| OfficeQA Pro | 53.2% | Coming soon | Coming soon |
| MMMU-Pro w/ Python | 82.1% | Coming soon | Coming soon |
| CharXiv | 82.8% | Coming soon | Coming soon |
| ERQA | 65.4% | Coming soon | Coming soon |
| SimpleVQA | 61.1% | Coming soon | Coming soon |
| ScreenSpot Pro | 85.4% | Coming soon | Coming soon |
| ZeroBench | 41.0% | Coming soon | Coming soon |
| MedXpertQA (MM) | 77.1% | Coming soon | Coming soon |

### Knowledge

- Winner: Directional only
- GPT-5.4 public-lane score: 68.6 (Supported · #22/189)
- LFM3-1.7B public-lane score: 25.5 (Estimated · #175/189)

| Benchmark | GPT-5.4 | LFM3-1.7B | Winner |
|-----------|-----------|-----------|--------|
| GPQA | 92.8% | 63.4% | GPT-5.4 |
| HLE | 52.1% | 7.2% | GPT-5.4 |
| HLE w/o tools | 39.8% | Coming soon | Coming soon |
| GPQA-D | 92.8% | 63.4% | GPT-5.4 |
| HealthBench Hard | 40.1% | Coming soon | Coming soon |
| MedXpertQA (Text) | 59.6% | Coming soon | Coming soon |
| HealthBench Professional | 48.1% | Coming soon | Coming soon |

### Instruction Following

- Winner: Directional only
- GPT-5.4 public-lane score: 90.7 (#25/156)
- LFM3-1.7B public-lane score: 71.7 (#76/156)

| Benchmark | GPT-5.4 | LFM3-1.7B | Winner |
|-----------|-----------|-----------|--------|
| IFEval | Coming soon | 96.7% | Coming soon |
| IFBench | Coming soon | 73.0% | Coming soon |
| Multi-IF | Coming soon | 89.10% | Coming soon |
| IFStruct | Coming soon | 93.72% | Coming soon |

### Mathematics

- Winner: Not comparable
- GPT-5.4 public-lane score: 82 (Unranked · 2 rankable rows)
- LFM3-1.7B public-lane score: Coming soon

| Benchmark | GPT-5.4 | LFM3-1.7B | Winner |
|-----------|-----------|-----------|--------|
| FrontierMath v2 (Tiers 1-3) | 78.596% | Coming soon | Coming soon |
| FrontierMath v2 (Tier 4) | 49.000% | Coming soon | Coming soon |

## FAQ

### Which is better overall, GPT-5.4 or LFM3-1.7B?

GPT-5.4 has the higher overall point estimate. Conditional score ranges do not establish rank confidence.

### Where is the biggest gap between GPT-5.4 and LFM3-1.7B?

The widest category gap is in knowledge, where the averages are 68.6 for GPT-5.4 and 25.5 for LFM3-1.7B.

### How many benchmarks does GPT-5.4 cover on BenchLM?

GPT-5.4 currently has 43 sourced benchmark scores on BenchLM.

### How many benchmarks does LFM3-1.7B cover on BenchLM?

LFM3-1.7B currently has 24 sourced benchmark scores on BenchLM.

## Related Comparisons

- [GPT-5.4 vs GPT-5.4 Pro](/compare/gpt-5-4-vs-gpt-5-4-pro)
- [LFM3-1.7B vs GPT-5.4 Pro](/compare/gpt-5-4-pro-vs-lfm3-1-7b)
- [GPT-5.4 vs GPT-5.4 mini](/compare/gpt-5-4-vs-gpt-5-4-mini)
- [LFM3-1.7B vs GPT-5.4 mini](/compare/gpt-5-4-mini-vs-lfm3-1-7b)

## Explore More

- [GPT-5.4 profile](/models/gpt-5-4)
- [LFM3-1.7B profile](/models/lfm3-1-7b)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)