Data as of October 29, 2026 · How the score is built
Released Aug 4, 2026128K contextLiquidAI LFM2.5-2.6B model card
LFM2.5-2.6B
Decision readingLFM2.5-2.6B scores 25.2 out of 100 and ranks #202 of 216. This profile shows 7 source-displayable benchmark rows; its strongest eligible category is Agentic at #123. Published weights can be self-hosted, but infrastructure cost varies and is not a comparable API token rate.
Released Aug 4, 2026 — see all recent releases
Decision snapshot
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Capability
25.2/100
field median 50.5#202 of 216 ranked models
Public
#202of 216
Verified —
Price
Self-hosted; infrastructure cost varies
input median $0.75No comparable first-party hosted token rate
Speed
197tok/s
field median 92 tok/sFirst token 12.02 s
Context
128Ktokens
field median 256,000Maximum output length is tracked separately
Strongest published evidence
Agentic ranks #123. Particularly useful for coding agents, browser research, and computer-use workflows.
Validate before choosing
7 published rows leave some tracked benchmark slots empty. No comparable first-party API token rate is published.
Source-linked · 7 displayable benchmark rows
Category score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #123 of 135Percentile 9thWeight 22%4 benchmarksVerified | 20.5 | 4 benchmarks | Verified | |||
| CodingRank #148 of 156Percentile 5thWeight 20%1 benchmarkVerified | 14.6 | 1 benchmark | Verified | |||
| ReasoningWeight 17%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MultimodalWeight 12%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| KnowledgeRank #160 of 179Percentile 11thWeight 12%Not listedUnlisted sources | 28.0 | Not listed | Unlisted sources | |||
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| Inst. FollowingRank #126 of 157Percentile 20thWeight 5%1 benchmarkVerified | 39.1 | 1 benchmark | Verified | |||
| MathWeight 5%1 benchmarkVerified | Score pending | 1 benchmark | Verified |
7 of 782 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic4/4 verified
- Coding1/1 verified
- ReasoningNot measured
- MultimodalNot measured
- KnowledgeNot measured
- MultilingualNot measured
- Inst. Following1/1 verified
- Math1/1 verified
Capability shape
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
LFM2.5-2.6B category percentile values
- Agentic9th percentile
- Coding5th percentile
- ReasoningNot eligible
- MultimodalNot eligible
- Knowledge11th percentile
- MultilingualNot eligible
- Instruction following20th percentile
- MathNot eligible
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#123/135
- Coding#148/156
- ReasoningNot ranked
- MultimodalNot ranked
- Knowledge#160/179
- MultilingualNot ranked
- Inst. Following#126/157
- MathNot ranked
Benchmark ledger
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
Coding1 row
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| LiveCodeBench v6 | Score59.4% | Versus best verified row Best verified: Sakana Fugu-Ultra · 93.2% | Gap33.8 behind | WeightDisplay only | Provider exact |
Agentic4 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| BFCL v4Berkeley Function Calling Leaderboard v4 | Score56.9% | Versus best verified row Best verified: BTL-3 · 88.5% | Gap31.6 behind | Weight3% ref. weight | Provider exact |
| τ³-bench resultsτ³-Bench Tool-Agent-User Evaluation | Score5.7% | Versus best verified row Best verified: Mercury 2.5 · 96.0% | Gap90.3 behind | Weight2% ref. weight | Provider exact |
| Claw-Eval | Score62.9% | Versus best verified row Best verified: Ornith-1.5-397B · 81.4% | Gap18.6 behind | WeightDisplay only | Provider exact |
| PinchBench | Score68.2% | Versus best verified row Best verified: Pokee-Isaac 28B · 95.7% | Gap27.5 behind | WeightDisplay only | Provider exact |
Inst. Following1 row
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| IFBenchInstruction Following Benchmark | Score59.2% | Versus best verified row Best verified: Flint · 89.2% | Gap30 behind | WeightWeighted 70% | Provider exact |
Math1 row
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| AIME 2025American Invitational Mathematics Examination 2025 | Score51.9% | Versus best verified row Best verified: GPT-5.2 · 100% | Gap48.1 behind | WeightDisplay only | Provider exact |
Bars run 0–100; the dark tick marks the best source-verified value
All 7 rowsLineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Reasoning
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- Not publishedLiquidAI Hugging Face model card
- Context window
- 128KLiquidAI Hugging Face model card
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- Not sourced yet
- Output modalities
- Not sourced yet
- Parameters
- Not sourced yet
- Availability
- Liquid AI publishes the 2.69B-parameter BF16 checkpoint and GGUF, ONNX, and MLX deployment formats under the LFM Open License v1.0. Liquid reports 220 output tokens/s on an Apple M5 Max and 113 tokens/s on an AMD Ryzen AI Max+ 395, with memory use under 2.5 GB; these are provider measurements, not BenchLM runtime tests.
- Cloud regions
- Not tracked yet
- Lifecycle
- Current
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Not documented in the pricing recordLiquidAI Hugging Face model card
- Self-host
- LFM2.5-2.6B needs ~24GB VRAM (1× NVIDIA RTX 4090 (24GB)).
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
LFM2.5-2.6B ranks #202 of 216 on the public leaderboard with a score of 25.19/100. It does not yet have enough sourced coverage for a verified position.
LFM2.5-2.6B is a open weight model with a 128K context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
Liquid AI publishes the 2.69B-parameter BF16 checkpoint and GGUF, ONNX, and MLX deployment formats under the LFM Open License v1.0. Liquid reports 220 output tokens/s on an Apple M5 Max and 113 tokens/s on an AMD Ryzen AI Max+ 395, with memory use under 2.5 GB; these are provider measurements, not BenchLM runtime tests.
7 of 782 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Agentic at #123, while its lowest eligible position is Knowledge at #160. particularly useful for coding agents, browser research, and computer-use workflows.
Last updated October 29, 2026. Runtime fields remain blank until a sourced snapshot exists.
Deployment options
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Questions
How does LFM2.5-2.6B perform overall in AI benchmarks?
LFM2.5-2.6B ranks #202 out of 216 models on the public BenchAlign leaderboard, with a score of 25.19/100. Its evidence status is Estimated, and this profile shows 7 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is LFM2.5-2.6B good for coding and programming?
LFM2.5-2.6B ranks #148 out of 156 eligible models for coding and programming, with a public category score of 14.6/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is LFM2.5-2.6B good for mathematics?
LFM2.5-2.6B has source-displayable benchmark coverage for mathematics, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is LFM2.5-2.6B good for agentic tool use and computer tasks?
LFM2.5-2.6B ranks #123 out of 135 eligible models for agentic tool use and computer tasks, with a public category score of 20.5/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is LFM2.5-2.6B good for instruction following?
LFM2.5-2.6B ranks #126 out of 157 eligible models for instruction following, with a public category score of 39.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is LFM2.5-2.6B open source?
LFM2.5-2.6B is an open-weight model from LiquidAI. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.
Does LFM2.5-2.6B have full benchmark coverage on BenchLM?
No. LFM2.5-2.6B currently has 16 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of LFM2.5-2.6B?
LFM2.5-2.6B has a documented context window of 128K. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.