Benchmarks of LFM2.5 2.6B12
BenchmarkByScoreRank
Overall1
Knowledge3
Science3
Agents2
Documents1
Legacy · saturated or no longer reported2
No benchmark matches
—per task?
free API onlyprice
—output speed?
131Kcontext
Rankingin thinking mode?
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-6 Lunalow×2.6—$0.005Function calling only with none · Chat Completions+3IOpenK2 Horizon 7Bthinking×2.5—$0.30 / $1.20Memory (4-bit): 1.3 GB → 4.5 GBBetterGPT-5.6 Lunalow×2.5—$0.01
BetterMistral Small 4thinkingOpen+34%—$0.015Memory (4-bit): 1.3 GB → 72 GB
BetterMiMo v2.5thinkingOpen×2.6—$0.019Memory (4-bit): 1.3 GB → 186 GBModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×2.6—$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.5—$0.01Mistral Small 4thinkingOpen+34%—$0.015Memory (4-bit): 1.3 GB → 72 GBMiMo v2.5thinkingOpen×2.6—$0.019Memory (4-bit): 1.3 GB → 186 GB
Granite 4.2 8BthinkingOpen+32%—$0.024Context: 131K → 128K+1ModelScoreCostPer taskWhat changes
IK2 Horizon 7BthinkingOpen×2.5—$0.30 / $1.20Memory (4-bit): 1.3 GB → 4.5 GBIK2 Horizon 3.7BthinkingOpen+86%—$0.30 / $1.20Memory (4-bit): 1.3 GB → 2.5 GB
Gemma 4 12BthinkingOpen+69%—$0.10 / $0.30Memory (4-bit): 1.3 GB → 6 GB
Qwen 3.5 9Bnon-thinkingOpen+58%—$0.08 / $0.13Memory (4-bit): 1.3 GB → 4.8 GBQwen3.5 4BthinkingOpen+56%—$0.03 / $0.15Memory (4-bit): 1.3 GB → 2.3 GBModelScoreCostPer taskWhat changes
lLFM2.5 24B A2BOpenn/a—$0.05 / $0.70Memory (4-bit): 1.3 GB → 14 GBAll liquid 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-11#98
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈8.4#241
AA Omniscience · Accuracy4.4#317
AA-LCR (long context) 1.15.7#270
GPQA Diamond55.8#245
Coding
SciCode14.4#139
Terminal-Bench 2.14.5#138
Agents & Tool Use
τ³-Banking · AA7.2#102
FAQ
How much does LFM2.5 2.6B cost?›
There is no per-unit API price: free API only.
Is LFM2.5 2.6B free to use?›
Yes, there is a free API option.
How good is LFM2.5 2.6B, and what is it best at?›
#242 of 334 in Chat & Knowledge, #139 of 140 in Coding, #122 of 175 in Agents & Tool Use.
Is LFM2.5 2.6B better than Nanbeige4.1 3B?›
On AA Intelligence Index 4.3.2, LFM2.5 2.6B scores 8.4 vs 8.4 — about the same.
What is the context window of LFM2.5 2.6B?›
131K tokens, with up to 8K output tokens.
When was LFM2.5 2.6B released? Is there a newer version?›
Released 2026-08-04. It is the current version in its line.
Is LFM2.5 2.6B open source?›
Yes, the weights are open (lfm1.0 licence).
Can I run LFM2.5 2.6B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can LFM2.5 2.6B read images?›
No, it accepts text only.
What is LFM2.5 2.6B?›
LFM2.5 2.6B is an open-weights language model from liquid, released Aug 2026.
What are the best alternatives to LFM2.5 2.6B?›
GPT-6 Luna (better), K2 Horizon 7B (open weights).