LFM2.5-8B-A1BTerbuka

Liquid AI · dirilis 28 Mei 2026

Liquid AI LFM2.5-8B-A1B model with 33K context window, optimized for advanced reasoning and multimodal tasks.

text33K ctx8.3B (1.5B active)
65.0
Podium Score
#213 dari 693
Intelligence Rank
65.0#213
AAQI: 30.0
Output Speed
338 t/s
TTFT: ~1.48s · ITL: 3.0 ms/tok
Blended Price (3:1)
$0.875 /1M
In: $0.5/M · Out: $2/M · Cache: $0.05
Context Capacity
33K
≈ 83 A4 Pages (Max Out: 4096)

Quality Domain Radar

6-Axis Frontier Evaluation vs Category Average & Flagship

LFM2.5-8B-A1B
Claude Mythos Preview (#1)
Average (68)
Coding (63)Mathematics (39)Hard Science (GPQA) (45)Agentic & Tool Use (36)Instruction Following (67)Vision / Multimodal (25)
API Deployment

Compare Multi-Host Hosting & Latency

Benchmark inference speed, TTFT latency distributions, and prompt cache pricing across Groq, Cerebras, Together, DeepInfra, Fireworks, AWS Bedrock, and Azure.

View API Providers Matrix →

Skor per sumber

Skor per kategori

Hasil benchmark

Detail model

Tanggal rilis28 Mei 2026
Parameter8.3B (1.5B active)
LisensiOpen Weights
Jendela konteks33,000 tokens
Harga input$0.5/M
Harga output$2/M
Kecepatan output338 t/s
Waktu ke token pertama1 ms

💰 Estimated Monthly Bill & TCO Calculator

Estimated Monthly Spend
$0.00
Calculating savings vs GPT-4o...

🔖 Embed Badges & Citation Snippet

GitHub Markdown Badge
[![LLMPodium Rank](https://llmpodium.com/badge/lfm2-5-8b-a1b/rank.svg)](https://llmpodium.com/models/lfm2-5-8b-a1b)
BibTeX Citation
@misc{llmpodium2026lfm258ba1b,
  title = {Benchmark Analysis and Intelligence Evaluation of LFM2.5-8B-A1B},
  author = {LLMPodium Research},
  year = {2026},
  url = {https://llmpodium.com/models/lfm2-5-8b-a1b}
}

Head-to-head

Model terbaik

model.alternatives_title

model.alternatives_desc

0 / 4 Models Selected