Skip to content
Model Pareto

LFM2.5-8B-A1B

Liquid AIOpen weightsartificialanalysis.ai ↗
Input
—
Output
—
Speed
~120tok/sest.
Context
33Ktokens

Long-tail import from Artificial Analysis (2026-09-24); AA variant: LFM2.5-8B-A1B. Price = AA median host price.

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs. No API sells this model, so this is its price on the chart.

Hardware
1× H200 141GB
BF16 weights
Throughput
~13,288 tok/s
many requests batched
Price
$0.031–0.26 /M tok
busy → light use
On your own machine
RTX 5090 32GB (int4, 300+ tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • BF16 weights (8.3B params, 1.5B active per token) + 25% KV-cache headroom ≈ 21 GB
  • 1× H200 141GB at $3.59–$7.91/GPU-hour on-demand (2026-09-24)
  • ~13,288 output tok/s aggregate at batch 256 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~155,768 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Writing & chat

MMMLUEstimated
~last of 5 (estimated)below measured range (<81.3%)
  1. #1Gemini 3.1 Pro92.6%
  2. #2Gemma 4 31B88.4%
  3. #4gpt-oss-120b81.3%
  4. ~#5LFM2.5-8B-A1BEstimated<81.3%
No published score — estimated from related benchmarks
~#46 of 57 (estimated)~1360 Elo
  1. #1Gemini 4 Argon1525
  2. #2Claude Opus 5.51507
  3. #45Trinity Large Thinking1369
  4. ~#46LFM2.5-8B-A1BEstimated1360
  5. #46INTELLECT-31356
No published score — estimated from related benchmarks
~#19 of 26 (estimated)~1566 Elo
  1. #1GPT-6 Astra2173
  2. #2Claude Fable 5.12162
  3. #18Inkling1611
  4. ~#19LFM2.5-8B-A1BEstimated1566
  5. #19Gemini 3.5 Flash-Lite1559
No published score — estimated from related benchmarks

Hard reasoning

Agents

~#12 of 15 (estimated)~52.9%
  1. #1Kimi K384.8%
  2. #2Qwen3.8 27B84.3%
  3. #11Muse Glimmer65.9%
  4. ~#12LFM2.5-8B-A1BEstimated52.9%
  5. #12Nemotron 3.5 Super VL NVFP452.3%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#3 of 5 (estimated)~53.8%
  1. #1GLM-5.3 Flash59.1%
  2. #2Kimi K358.3%
  3. ~#3LFM2.5-8B-A1BEstimated53.8%
  4. #3GPT-5.6 Terra50.2%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#21 of 22 (estimated)~35.4%
  1. #1Atria Dawn Preview92.5%
  2. #2GPT-6 Astra91.5%
  3. #20Nemotron 3 Ultra44.4%
  4. ~#21LFM2.5-8B-A1BEstimated35.4%
  5. #21Mistral Small 421.3%
No published score — estimated from related benchmarks
~#60 of 92 (estimated)~716 Elo
  1. #1Claude Opus 5.51866
  2. #2Muse Wonder Full1857
  3. #59Claude Haiku 4.5719
  4. ~#60LFM2.5-8B-A1BEstimated716
  5. #60MiniCPM5-2B694
No published score — estimated from related benchmarks
~#79 of 85 (estimated)~0.3%
  1. #1Claude Fable 5.276.4%
  2. #2Muse Wonder Full71.3%
  3. #77Granite 4.2 3B0.4%
  4. ~#79LFM2.5-8B-A1BEstimated0.3%
  5. #79gpt-oss-120b0.2%
No published score — estimated from related benchmarks
~#81 of 91 (estimated)~5.4%
  1. #1Muse Spark 1.350.5%
  2. #2GLM-5.350.3%
  3. #80Qwen3 Coder Next5.4%
  4. ~#81LFM2.5-8B-A1BEstimated5.4%
  5. #81K2 Horizon 0.9B4.9%
No published score — estimated from related benchmarks