Skip to main content

Model profile

LFM2.5-8B-A1B

LiquidAICurrentReleased May 28, 2026
Data verified
Overall Score
41.42Public #166 of 200
Arena Elo
Not listed
Eligible category ranks
2of 8
Price (1M tokens)
$0 in / $0 out
API pricing
Speed
Not listed
Context
128K

Evidence coverage

17 of 323 tracked benchmarks are published. 9 are verified and 8 provisional. 6 of 8 categories are measured.

Updated July 23, 2026Methodology
Published / tracked
17 / 323
Verified
9
Provisional
8
Categories with evidence
6 / 8

Evidence by category

  • Agentic2 benchmarks
    Verified
  • Coding1 benchmark
    Reported
  • Reasoning2 benchmarks
    Reported
  • Knowledge6 benchmarks
    Mixed evidence
  • Math3 benchmarks
    Verified
  • Multilingual0 benchmarks
    Not measured
  • Multimodal0 benchmarks
    Not measured
  • Inst. Following3 benchmarks
    Mixed evidence
Open WeightSelf-hostReasoning
Confidence:
Low
reasoning

LFM2.5-8B-A1B ranks #166 out of 200 models on the public leaderboard with an overall score of 41.42/100. It does not yet have enough sourced coverage for BenchLM's verified leaderboard. While not a frontier model, it offers specific advantages depending on the use case.

LFM2.5-8B-A1B is a open weight model with a 128K token context window. It uses explicit chain-of-thought reasoning, which typically improves performance on math and complex reasoning tasks at the cost of higher latency and token usage.

This profile currently has 17 of 323 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Instruction Following (#24), while its weakest is Agentic (#68). This performance profile makes it a well-rounded choice across a range of tasks.

Peer position

Exact provisional scores and ranks for the closest listed peers. A score can appear before a model clears the evidence threshold for a rank, so equal scores can have different rank states.

Range 41.1342.06

  1. GPT-4.1 nano
    OpenAI
    #16142.06
    GPT-4.1 nano is #161 with a score of 42.06.
    Compare
  2. Gemma 4 E2B
    Google
    #16241.82
    Gemma 4 E2B is #162 with a score of 41.82.
    Compare
  3. Mistral Large 2
    Mistral
    #16341.77
    Mistral Large 2 is #163 with a score of 41.77.
    Compare
  4. Gemma 3 27B
    Google
    #16441.57
    Gemma 3 27B is #164 with a score of 41.57.
    Compare
  5. GPT-4o
    OpenAI
    #16541.49
    GPT-4o is #165 with a score of 41.49.
    Compare
  6. LFM2.5-8B-A1BCurrent model
    LiquidAI
    #16641.42
    LFM2.5-8B-A1B is #166 with a score of 41.42.
  7. Claude 3 Opus
    Anthropic
    #16841.13
    Claude 3 Opus is #168 with a score of 41.13.
    Compare

Category percentile

More

Relative position among models eligible for each sourced category. A higher percentile means a stronger position within that category's ranked cohort; 100 is highest.

  1. Inst. Following23%
    Eligible cohort rank #24 of 31Category score 55.7
  2. Agentic43%
    Eligible cohort rank #68 of 119Category score 46.0

Category evidence

Scores and ranks appear only where this model has published benchmark evidence. Categories without displayable source records remain not measured.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #68 of 119Percentile 43rdWeight 22%2 benchmarksVerified46.0
CodingWeight 20%1 benchmarkReportedScore pending
ReasoningWeight 17%2 benchmarksReportedScore pending
KnowledgeWeight 12%6 benchmarksMixed sourcesScore pending
MathRank Not rankedWeight 5%3 benchmarksVerified27.8
MultilingualWeight 7%0 benchmarksNot measuredNot measured
MultimodalWeight 12%0 benchmarksNot measuredNot measured
Inst. FollowingRank #24 of 31Percentile 23rdWeight 5%3 benchmarksMixed sources55.7

Benchmark Details

Rows below have a displayable published verification record. Each source link and provenance note remains in the page HTML while its category is closed. Source-unverified manual rows and generated rows stay hidden.

Agentic2 benchmarks
BFCL v4Provider exact

Berkeley Function Calling Leaderboard v4

49.7%Display only
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 49.73 on BFCLv4.
τ²-bench resultsProvider exact

τ²-Bench Tool-Agent-User Evaluation

16.1%Display only
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 88.07 on Tau² Telecom. BenchLM stores it on the existing tau2Bench key.
Coding1 benchmark
AA-SciCodeReported

Artificial Analysis SciCode

7.8%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Reasoning2 benchmarks
AA-LCRReported

Artificial Analysis Long Context Reasoning

0.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
CritPtReported

Critical Physics Tasks

0.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Knowledge6 benchmarks
AA-GPQA DiamondReported

Artificial Analysis GPQA Diamond

51.3%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-HLEReported

Artificial Analysis Humanity's Last Exam

6.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience IndexProvider exact

Artificial Analysis Omniscience Index

-33.3%Display only
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at -24.70 on AA-Omniscience Index.
AA-Omniscience AccuracyProvider exact

Artificial Analysis Omniscience Accuracy

9.4%Display only
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 8.67 on AA-Omniscience Accuracy.
AA-Omniscience Hallucination RateReported

Artificial Analysis Omniscience Hallucination Rate

47.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Artificial Analysis Intelligence IndexReported
8.3%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Math3 benchmarks
AIME26Provider exact

AIME 2026

50.0%Weighted 25%
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 50.00 on AIME26.
MATH-500Provider exact

MATH-500 Problem Set

88.8%Display only
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 88.76 on MATH500.
AIME 2025Provider exact

American Invitational Mathematics Examination 2025

42.5%Display only
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 42.53 on AIME25.
Inst. Following3 benchmarks
IFBenchProvider exact

Instruction Following Benchmark

56.5%Weighted 65%
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 56.47 on IFBench.
IFEvalProvider exact

Instruction-Following Eval

91.8%Weighted 35%
Source: Liquid AI: LFM2.5-8B-A1B launch postProvenance: Liquid reports LFM2.5-8B-A1B at 91.84 on IFEval.
AA-IFBenchReported

Artificial Analysis IFBench

55.6%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.

Frequently Asked Questions

How does LFM2.5-8B-A1B perform overall in AI benchmarks?

LFM2.5-8B-A1B has 17 published benchmark scores on BenchLM, but it does not yet have enough non-generated coverage to receive a global overall rank.

Is LFM2.5-8B-A1B good for knowledge and understanding?

LFM2.5-8B-A1B has visible benchmark coverage in knowledge and understanding, but BenchLM does not currently assign it a global category rank there.

Is LFM2.5-8B-A1B good for coding and programming?

LFM2.5-8B-A1B has visible benchmark coverage in coding and programming, but BenchLM does not currently assign it a global category rank there.

Is LFM2.5-8B-A1B good for mathematics?

LFM2.5-8B-A1B has visible benchmark coverage in mathematics, but BenchLM does not currently assign it a global category rank there.

Is LFM2.5-8B-A1B good for reasoning and logic?

LFM2.5-8B-A1B has visible benchmark coverage in reasoning and logic, but BenchLM does not currently assign it a global category rank there.

Is LFM2.5-8B-A1B good for agentic tool use and computer tasks?

LFM2.5-8B-A1B ranks #68 out of 119 models in agentic tool use and computer tasks benchmarks with an average score of 46. There are stronger options in this category.

Is LFM2.5-8B-A1B good for instruction following?

LFM2.5-8B-A1B ranks #24 out of 31 models in instruction following benchmarks with an average score of 55.7. There are stronger options in this category.

Is LFM2.5-8B-A1B open source?

Yes, LFM2.5-8B-A1B is an open weight model created by LiquidAI, meaning it can be downloaded and run locally or fine-tuned for specific use cases.

Does LFM2.5-8B-A1B have full benchmark coverage on BenchLM?

Not yet. LFM2.5-8B-A1B currently has 17 published benchmark scores out of the 323 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of LFM2.5-8B-A1B?

LFM2.5-8B-A1B has a published context window of 128K, which determines how much text it can process in a single interaction.

Last updated: July 23, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Choose with this week’s evidence

Join 2,000+ readers for ranking moves, new releases, pricing changes, and the evidence behind them.

Free. One email per week.