Skip to content
Model Pareto

LFM2.5-24B-A2B

Liquid AIOpen weightshuggingface.co ↗
Compare with nearest rivalsCompare with…
Input
—
Output
—
Speed
~110tok/sest.
Context
131Ktokens

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
~7,947 tok/s
many requests batched
Price
$0.052–0.43 /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 (24B params, 2.45B active per token) + 25% KV-cache headroom ≈ 60 GB
  • 1× H200 141GB at $3.59–$7.91/GPU-hour on-demand (2026-09-24)
  • ~7,947 output tok/s aggregate at batch 256 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~95,368 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.

Coding

SWE-bench Pro (Public)Estimated
~last of 23 (estimated)below measured range (<43.8%)
  1. #1Claude Opus 5.589.9%
  2. #2Claude Sonnet 5.581.3%
  3. #22Nex-N2.5-Mini43.8%
  4. ~#23LFM2.5-24B-A2BEstimated<43.8%
No published score — estimated from related benchmarks
DeepSWE v1.1Estimated
~#40 of 43 (estimated)~29.4%
  1. #1Gemini 4 Argon78.8%
  2. #2DeepSeek V4.1 Flash 101777.0%
  3. #39Nex-N2.5-Mini36.1%
  4. ~#40LFM2.5-24B-A2BEstimated29.4%
  5. #40Hy328.0%
No published score — estimated from related benchmarks
Terminal-Bench 4.0Estimated
~#39 of 99 (estimated)~4.2%
  1. #1Kimi K3.1 Pro67.2%
  2. #2Claude Sonnet 5.563.6%
  3. #38Qwen3.8 27B5.6%
  4. ~#39LFM2.5-24B-A2BEstimated4.2%
  5. #39Gemini 3.1 Pro4.0%
No published score — estimated from related benchmarks
LiveCodeBenchEstimated
~#17 of 28 (estimated)~59.8%
  1. #1Qwen3.8-Omni-Flash92.6%
  2. #2Qwen3.8 Flash-Next91.9%
  3. #16Claude Haiku 4.561.5%
  4. ~#17LFM2.5-24B-A2BEstimated59.8%
  5. #17GLM-5.3 NVFP453.5%
No published score — estimated from related benchmarks
SciCodeEstimated
~#93 of 118 (estimated)~36.4%
  1. #1Claude Opus 5.566.9%
  2. #2Claude Fable 5.163.1%
  3. #91Qwen3.6 35B A3B36.6%
  4. ~#93LFM2.5-24B-A2BEstimated36.4%
  5. #93LongCat 2.036.3%
No published score — estimated from related benchmarks
AA Coding Agent IndexEstimated
~#22 of 24 (estimated)~31.8
  1. #1Claude Opus 5.561.7
  2. #2Claude Opus 559.7
  3. #21Gemini 3.8 Flash33.3
  4. ~#22LFM2.5-24B-A2BEstimated31.8
  5. #22GPT-6 Luna31.1
No published score — estimated from related benchmarks

Writing & chat

MMMLUEstimated
~#3 of 5 (estimated)~84.7%
  1. #1Gemini 3.1 Pro92.6%
  2. #2Gemma 4 31B88.4%
  3. ~#3LFM2.5-24B-A2BEstimated84.7%
  4. #3Claude Haiku 4.583.0%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Estimated
~#44 of 57 (estimated)~1374 Elo
  1. #1Gemini 4 ArgonFew votes1512
  2. #2Claude Opus 5.51497
  3. #43Qwen3 Next 80B A3B Instruct1386
  4. ~#44LFM2.5-24B-A2BEstimated1374
  5. #44GLM-4.6VFew votes1361
No published score — estimated from related benchmarks
LMArena Text (Elo)Estimated
~#44 of 57 (estimated)~1393 Elo
  1. #1Gemini 4 Argon1525
  2. #2Claude Opus 5.51507
  3. #43Qwen3 Next 80B A3B Instruct1399
  4. ~#44LFM2.5-24B-A2BEstimated1393
  5. #44GLM-4.6VFew votes1378
No published score — estimated from related benchmarks
EQ-Bench Creative Writing v3 (Elo)Estimated
~#18 of 26 (estimated)~1638 Elo
  1. #1GPT-6 Astra2173
  2. #2Claude Fable 5.12162
  3. #17Qwen3.8 27B1671
  4. ~#18LFM2.5-24B-A2BEstimated1638
  5. #18Inkling1611
No published score — estimated from related benchmarks

Hard reasoning

AIME (latest)Estimated
~#22 of 33 (estimated)~66.4%
  1. #1GLM-5.3 NVFP499.5%
  2. #2Beam 1.0 501B A23B97.7%
  3. #21Phi-4-reasoning-plus78.0%
  4. ~#22LFM2.5-24B-A2BEstimated66.4%
  5. #22Mistral Large 338.0%
No published score — estimated from related benchmarks