Model comparison
GPT-5.4 vs LFM2.5-8B-A1B
Head-to-head evidence from 12 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Verified leaderboard positions: GPT-5.4 #10; LFM2.5-8B-A1B unranked
BenchAlign evidence: GPT-5.4 supported; LFM2.5-8B-A1B estimated. Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.4 and LFM2.5-8B-A1B share 12 comparable benchmark results. 1 of 8 categories are comparable. 42 results are unique to GPT-5.4; 6 to LFM2.5-8B-A1B.
Updated July 16, 2026- Shared results
- 12
- GPT-5.4 only
- 42
- LFM2.5-8B-A1B only
- 6
- Comparable categories
- 1 / 8
Pick GPT-5.4 if you want the stronger benchmark profile. LFM2.5-8B-A1B only becomes the better choice if mathematics is the priority or you want the cheaper token bill.
Confidence note. This is a partial-evidence comparison with 12 shared benchmark results across 5 evidence categories; 1 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
GPT-5.4 is clearly ahead on the provisional aggregate, 86 to 37. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.4 is also the more expensive model on tokens at $2.50 input / $15.00 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for LFM2.5-8B-A1B. That is roughly Infinityx on output cost alone. GPT-5.4 gives you the larger context window at 1.05M, compared with 128K for LFM2.5-8B-A1B.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | GPT-5.4 | Δ | LFM2.5-8B-A1B |
|---|---|---|---|
| Math | GPT-5.442.5 | Margin→ 7.5 | LFM2.5-8B-A1B50.0 |
| Agentic | GPT-5.477.2 | MarginNo overlap | LFM2.5-8B-A1BNot measured |
| Coding | GPT-5.457.7 | MarginNo overlap | LFM2.5-8B-A1BNot measured |
| Knowledge | GPT-5.457.6 | MarginNo overlap | LFM2.5-8B-A1BNot measured |
| Multimodal | GPT-5.473.2 | MarginNo overlap | LFM2.5-8B-A1BNot measured |
| Inst. Following | GPT-5.4Not measured | MarginNo overlap | LFM2.5-8B-A1B68.8 |
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.4 | LFM2.5-8B-A1B | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.4$2.5 input / $15 output | LFM2.5-8B-A1B$0 input / $0 output | LFM2.5-8B-A1B has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.474 tok/s | LFM2.5-8B-A1BNot available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | GPT-5.4151.79 s | LFM2.5-8B-A1BNot available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | GPT-5.41.05M | LFM2.5-8B-A1B128K | GPT-5.4 lists the larger context window. |
Benchmark Deep Dive
Agentic18 benchmarks
| Benchmark | GPT-5.4 | LFM2.5-8B-A1B | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 75.1% | — | Not comparable |
| CyberGymSource | 79.0% | — | Not comparable |
| BrowseCompSource | 82.7% | — | Not comparable |
| OSWorld-VerifiedSource | 75% | — | Not comparable |
| MCP AtlasSource | 70.6% | — | Not comparable |
| ToolathlonSource | 54.6% | — | Not comparable |
| τ²-bench resultsSource | 87.1% | 16.1% | GPT-5.4 leads |
| Claw-EvalSource | 60.3% | — | Not comparable |
| DeepSearchQASource | 73.6% | — | Not comparable |
| AA Agentic IndexSource | 41.1% | — | Not comparable |
| APEX-Agents-AASource | 33.3% | — | Not comparable |
| GDPval-AASource | 44.7% | — | Not comparable |
| GDPval-AASource | 1395 | — | Not comparable |
| Gert LabsSource | 64.89% | — | Not comparable |
| ResearchClawBenchSource | 15.3% | — | Not comparable |
| JobBenchSource | 38.9% | — | Not comparable |
| ExploitGymSource | 6.0% | — | Not comparable |
| BFCL v4Source | — | 49.7% | Not comparable |
Coding7 benchmarks
| Benchmark | GPT-5.4 | LFM2.5-8B-A1B | Result |
|---|---|---|---|
| LiveCodeBench ProSource | 87.5% | — | Not comparable |
| SWE-bench ProSource | 57.7% | — | Not comparable |
| React Native EvalsSource | 85.3% | — | Not comparable |
| Vibe Code BenchSource | 67.42% | — | Not comparable |
| AA Coding IndexSource | 71.0% | — | Not comparable |
| Terminal-Bench HardSource | 57.6% | 4.5% | GPT-5.4 leads |
| AA-SciCodeSource | 56.6% | 7.8% | GPT-5.4 leads |
Reasoning2 benchmarks
Knowledge13 benchmarks
| Benchmark | GPT-5.4 | LFM2.5-8B-A1B | Result |
|---|---|---|---|
| GPQASource | 92.8% | — | Not comparable |
| HLESource | 52.1% | — | Not comparable |
| HLE w/o toolsSource | 39.8% | — | Not comparable |
| GPQA-DSource | 92.8% | — | Not comparable |
| HealthBench HardSource | 40.1% | — | Not comparable |
| MedXpertQA (Text)Source | 59.6% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 51.4% | 8.3% | GPT-5.4 leads |
| AA-GPQA DiamondSource | 92.0% | 51.3% | GPT-5.4 leads |
| AA-HLESource | 41.6% | 6.9% | GPT-5.4 leads |
| AA-Omniscience IndexSource | 5.7% | -33.3% | GPT-5.4 leads |
| AA-Omniscience AccuracySource | 50.0% | 9.4% | GPT-5.4 leads |
| AA-Omniscience Hallucination RateSource | 88.6% | 47.0% | LFM2.5-8B-A1B leads |
| HealthBench ProfessionalSource | 48.1% | — | Not comparable |
MathLFM2.5-8B-A1B wins5 benchmarks
Multimodal12 benchmarks
| Benchmark | GPT-5.4 | LFM2.5-8B-A1B | Result |
|---|---|---|---|
| MMMU-ProSource | 81.2% | — | Not comparable |
| OfficeQA ProSource | 53.2% | — | Not comparable |
| MMMU-Pro w/ PythonSource | 82.1% | — | Not comparable |
| CharXivSource | 82.8% | — | Not comparable |
| ERQASource | 65.4% | — | Not comparable |
| SimpleVQASource | 61.1% | — | Not comparable |
| ScreenSpot ProSource | 85.4% | — | Not comparable |
| ZeroBenchSource | 41.0% | — | Not comparable |
| MedXpertQA (MM)Source | 77.1% | — | Not comparable |
| GDPval-AASource | 1672 | — | Not comparable |
| AA-MMMU-ProSource | 78.4% | — | Not comparable |
| Design Arena WebsiteSource | 1254 | — | Not comparable |
Frequently Asked Questions (2)
Which is better, GPT-5.4 or LFM2.5-8B-A1B?
GPT-5.4 is ahead on BenchLM's provisional leaderboard, 86 to 37.
Which is better for math, GPT-5.4 or LFM2.5-8B-A1B?
LFM2.5-8B-A1B has the edge for math in this comparison, averaging 50 versus 42.5. GPT-5.4 stays close enough that the answer can still flip depending on your workload.
Related Comparisons
Explore More
The AI models change fast. We track them for you.
A weekly brief for engineers and researchers covering new models, ranking shifts, and pricing changes.
Free. No spam. Unsubscribe anytime.