LFM2.5 1.2B Thinking
Jan 2026
LFM2.5 1.2B ThinkingCompare inon
Benchmarks of LFM2.5 1.2B Thinking10
BenchmarkByScoreRank
Overall1
Knowledge4
Science2
Coding1
Legacy · saturated or no longer reported2
No benchmark matches
—per task?
self-host onlyprice
—output speed?
33Kcontext
Rankingin thinking mode?
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-6 Lunalow×4.1—$0.005Function calling only with none · Chat Completions+3IOpenK2 Horizon 7Bthinking×3.9—$0.30 / $1.20Memory (4-bit): 0.6 GB → 4.5 GB
BetterGranite 4.2 3BthinkingOpen+73%—$0.006Memory (4-bit): 0.6 GB → 1.8 GB
BetterGemma 4 E4BthinkingOpen+71%—$0.008Memory (4-bit): 0.6 GB → 4 GBBetterGPT-5.6 Lunalow×4—$0.01ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×4.1—$0.005Function calling only with none · Chat Completions+3Granite 4.2 3BthinkingOpen+73%—$0.006Memory (4-bit): 0.6 GB → 1.8 GBGemma 4 E4BthinkingOpen+71%—$0.008Memory (4-bit): 0.6 GB → 4 GBGPT-5.6 Lunalow×4—$0.01GPT OSS 20BhighOpen+72%—$0.011Memory (4-bit): 0.6 GB → 13 GBModelScoreCostPer taskWhat changes
IK2 Horizon 7BthinkingOpen×3.9—$0.30 / $1.20Memory (4-bit): 0.6 GB → 4.5 GBIK2 Horizon 3.7BthinkingOpen×3—$0.30 / $1.20Memory (4-bit): 0.6 GB → 2.5 GBGemma 4 12BthinkingOpen×2.7—$0.10 / $0.30Memory (4-bit): 0.6 GB → 6 GB
Qwen 3.5 9Bnon-thinkingOpen×2.5—$0.08 / $0.13Memory (4-bit): 0.6 GB → 4.8 GBQwen3.5 4BthinkingOpen×2.5—$0.03 / $0.15Memory (4-bit): 0.6 GB → 2.3 GBModelScoreCostPer taskWhat changes
LFM2.5 8B A1BthinkingOpen+38%——Memory (4-bit): 0.6 GB → 4.2 GBLFM2 24B A2Bnon-thinkingOpensame—$0.03 / $0.12Memory (4-bit): 0.6 GB → 14 GBLFM2 2.6BOpensame——Memory (4-bit): 0.6 GB → 1.3 GBLFM2.5 1.2B Instructnon-thinkingOpensame——LFM2 8B A1BOpensame——Memory (4-bit): 0.6 GB → 4.2 GBLFM2.5 VL 1.6Bnon-thinkingOpensame——Memory (4-bit): 0.6 GB → 0.8 GBLFM2 1.2BOpensame——All Liquid AI models →Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · full quality✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
Licencelfm1.0commercial use with conditions
Benchmarks
all categoriesChat & Knowledge
Humanity's Last Exam6.2#244
AA Omniscience Index-80#314
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈5.2#314
AA Omniscience · Accuracy7.4#299
GPQA Diamond33.9#290
Coding
Agents & Tool Use
τ²-bench · AA19.6#196
Writing & Language
IFBench (instructions)41.8#146
FAQ
How much does LFM2.5 1.2B Thinking cost?›
There is no per-unit API price: self-host only.
Is LFM2.5 1.2B Thinking free to use?›
There is no free API. The weights are open, so you can run it yourself.
How good is LFM2.5 1.2B Thinking, and what is it best at?›
#242 of 334 in Chat & Knowledge, #210 of 239 in Coding, #194 of 242 in Agents & Tool Use, #145 of 245 in Writing & Language.
Is LFM2.5 1.2B Thinking better than Olmo 3 7B Instruct?›
On AA Intelligence Index 4.3.2, LFM2.5 1.2B Thinking scores 5.2 vs 5.2 — about the same.
What is the context window of LFM2.5 1.2B Thinking?›
33K tokens.
When was LFM2.5 1.2B Thinking released? Is there a newer version?›
Released 2026-01-20. It is the current version in its line.
Is LFM2.5 1.2B Thinking open source?›
Yes, the weights are open (lfm1.0 licence).
Can I run LFM2.5 1.2B Thinking locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can LFM2.5 1.2B Thinking read images?›
No, it accepts text only.
What is LFM2.5 1.2B Thinking?›
LFM2.5 1.2B Thinking is an open-weights language model from Liquid AI, released Jan 2026.
What are the best alternatives to LFM2.5 1.2B Thinking?›
GPT-6 Luna (better), K2 Horizon 7B (open weights).