Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
LFM2.5-350M
~39
Winner · 0/8 categoriesMixtral 8x22B Instruct v0.1
36
1/8 categoriesLFM2.5-350M· Mixtral 8x22B Instruct v0.1
Pick LFM2.5-350M if you want the stronger benchmark profile. Mixtral 8x22B Instruct v0.1 only becomes the better choice if knowledge is the priority or you need the larger 64K context window.
LFM2.5-350M has the cleaner overall profile here, landing at 39 versus 36. It is a real lead, but still close enough that category-level strengths matter more than the headline number.
Mixtral 8x22B Instruct v0.1 gives you the larger context window at 64K, compared with 32K for LFM2.5-350M.
BenchLM keeps the benchmark table and the operator tradeoffs on the same page so a better score does not hide a materially slower, pricier, or smaller-context model.
Runtime metrics show N/A when BenchLM does not have a sourced snapshot for that exact model. The scoring rules and freshness policy are documented on the methodology page.
| Benchmark | LFM2.5-350M | Mixtral 8x22B Instruct v0.1 |
|---|---|---|
| Agentic | ||
| Terminal-Bench 2.0 | — | 35% |
| BrowseComp | — | 32% |
| OSWorld-Verified | — | 28% |
| Coding | ||
| HumanEval | — | 54.8% |
| SWE-bench Pro | — | 40% |
| Multimodal & Grounded | ||
| MMMU-Pro | — | 35% |
| OfficeQA Pro | — | 36% |
| Reasoning | ||
| LongBench v2 | — | 39% |
| MRCRv2 | — | 38% |
| KnowledgeMixtral 8x22B Instruct v0.1 wins | ||
| GPQA | 30.6% | — |
| MMLU-Pro | 20.0% | — |
| MMLU | — | 77.8% |
| FrontierScience | — | 53% |
| Instruction Following | ||
| IFEval | 77.0% | — |
| Multilingual | ||
| MMLU-ProX | — | 42% |
| Mathematics | ||
| Coming soon | ||
LFM2.5-350M is ahead overall, 39 to 36.
Mixtral 8x22B Instruct v0.1 has the edge for knowledge tasks in this comparison, averaging 53 versus 23.8. LFM2.5-350M stays close enough that the answer can still flip depending on your workload.
Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.