Poolside
Laguna-S-2.1
What Laguna-S-2.1’s own documents say it is built from — every method with the sentence that describes it.
- Parameters
- 118B / 8B active
- Context
- 1M
- Attention
- sliding/global hybrid1:3 global : sliding-window (512) GQA, per-head softplus output gating
- Experts
- 256 routed + 1 shared, top-10
- Input
- text
- Licence
- OpenMDW-1.1
- Released
- 2026-07-21
Against the consensus recipe
Its documents state 5 of the 18 methods the field agrees on.
Data curation
- Benchmark decontaminationnot in its documents
- Synthetic data generationnot in its documents
- Corpus deduplicationnot in its documents
Training objective
- Multi-Token Predictionnot in its documents
- Knowledge insulationnot in its documents
- Quantization-aware distillationused
Model architecture
- Mixture of expertscore
- Multi-Token Predictionnot in its documents
- Grouped-query attentioncore
Optimization
- Muoncore
- Expert parallelismnot in its documents
- Quantization-Aware Trainingnot in its documents
Post-training
- Supervised fine-tuningnot in its documents
- Asynchronous reinforcement learningnot in its documents
- Multi-Teacher On-Policy Distillationnot in its documents
Inference & serving
- Thinking modedefault
- Multi-Token Predictionnot in its documents
- FP8not in its documents
Checked in its code
Its code shows 11 of the 19 architecture features checked — 7 of them are not stated in its documents.
Sources say what a lab chose to describe; the modeling code says what the checkpoint runs. Each feature below was put to a code verifier, and only the config values and source lines it cited that were found in the files are shown, linked to the line.
Read from Hugging Face Transformers (modeling_laguna.py, configuration_laguna.py) and config.json at revision 0f57314083, by anthropic/claude-sonnet-5 on 2026-11-07.
| Feature | Code | Its documents | Verdict |
|---|---|---|---|
| Grouped-query attention | in its code | adopted | Each layer projects queries at num_attention_heads_per_layer (48 or 72) while K/V use num_key_value_heads=8, shared via repeat_kv groups. show evidencenum_key_value_heads = 8num_attention_heads = 48 |
| Multi-head latent attention | not in its code | not stated | There is no kv_lora_rank/q_lora_rank low-rank latent projection; attention uses standard per-head K/V projections sized by… show the whole verdictThere is no kv_lora_rank/q_lora_rank low-rank latent projection; attention uses standard per-head K/V projections sized by num_key_value_heads. |
| Sliding-window attention code only | in its code | not stated | Layers whose layer_types entry is sliding_attention set self.sliding_window = config.sliding_window (512), restricting attention to a… show evidenceLayers whose layer_types entry is sliding_attention set self.sliding_window = config.sliding_window (512), restricting attention to a local window via create_sliding_window_causal_mask. sliding_window = 512layer_types = ["full_attention", "sliding_attention", "sliding_attention", "sliding_attenti… |
| Interleaved sliding-window and global attention | in its code | adopted | layer_types fixes a pattern of one full_attention layer followed by three sliding_attention layers, repeated across all 48 layers. show evidencelayer_types = ["full_attention", "sliding_attention", "sliding_attention", "sliding_attenti… |
| Indexer-selected sparse attention (DeepSeek Sparse Attention) | not in its code | not stated | No indexer scoring module or index_topk/index_n_heads config keys exist; attention is dense GQA/sliding-window softmax attention only. |
| Linear-attention or state-space layers alongside full attention | not in its code | not stated | All token-mixing layers in LagunaDecoderLayer use LagunaAttention (softmax attention); no linear-attention or SSM layer type exists in… show the whole verdictAll token-mixing layers in LagunaDecoderLayer use LagunaAttention (softmax attention); no linear-attention or SSM layer type exists in layer_types or the code. |
| Gated DeltaNet layers | not in its code | not stated | No gated delta-rule recurrent mechanism or linear_num_key_heads/linear_conv_kernel_dim config exists in this model. |
| Mamba-2 layers | not in its code | not stated | No Mamba-2 SSM block or mamba_num_heads/ssm_state_size config exists; all layers are LagunaAttention. |
| Learnable attention sink | not in its code | not stated | The attention forward and eager_attention_forward implementation have no learned per-head sink logit added to the softmax denominator. |
| Gated attention output code only | in its code | not stated | Attention output is multiplied by a softplus (not sigmoid, but input-dependent) gate computed from g_proj before the output projection… show evidenceAttention output is multiplied by a softplus (not sigmoid, but input-dependent) gate computed from g_proj before the output projection, per-head as configured by gating='per-head'. gating = "per-head" |
| QK normalization code only | in its code | not stated | Queries and keys are each passed through a per-head RMSNorm (q_norm/k_norm) before rotary embedding and the attention dot product. show evidence |
| Partial RoPE code only | in its code | not stated | full_attention layers use partial_rotary_factor 0.5 (rotary applied to half of head_dim), while sliding_attention layers use 1.0 (full… show evidencefull_attention layers use partial_rotary_factor 0.5 (rotary applied to half of head_dim), while sliding_attention layers use 1.0 (full rotary). rope_parameters.full_attention.partial_rotary_factor = 0.5rope_parameters.sliding_attention.partial_rotary_factor = 1.0 |
| YaRN RoPE scaling code only | in its code | not stated | The full_attention rope_parameters entry sets rope_type yarn with factor/beta_fast/beta_slow, and the rotary embedding dispatches to… show evidenceThe full_attention rope_parameters entry sets rope_type yarn with factor/beta_fast/beta_slow, and the rotary embedding dispatches to ROPE_INIT_FUNCTIONS for non-default types. rope_parameters.full_attention.rope_type = "yarn"rope_parameters.full_attention.factor = 128.0 |
| Layers without positional encoding (NoPE) | not in its code | not stated | Every layer type (full_attention and sliding_attention) has a rope_parameters entry with a defined rope_type, so no softmax attention… show evidenceEvery layer type (full_attention and sliding_attention) has a rope_parameters entry with a defined rope_type, so no softmax attention layer skips positional encoding entirely. rope_parameters.sliding_attention.rope_type = "default" |
| Mixture of experts | in its code | adopted | LagunaSparseMoeBlock routes tokens via LagunaTopKRouter top-k=10 over num_experts=256 experts for layers marked 'sparse' in mlp_layer_types. show evidencenum_experts = 256num_experts_per_tok = 10 |
| Shared expert | in its code | adopted | Each sparse MoE block includes a shared_experts LagunaMLP with intermediate size shared_expert_intermediate_size=1024 added to every… show evidenceEach sparse MoE block includes a shared_experts LagunaMLP with intermediate size shared_expert_intermediate_size=1024 added to every token's output. shared_expert_intermediate_size = 1024 |
| Auxiliary-loss-free load balancing (selection bias) code only | in its code | not stated | The router adds a learned e_score_correction_bias to routing scores only for top-k selection, not for weighting, matching… show evidenceThe router adds a learned e_score_correction_bias to routing scores only for top-k selection, not for weighting, matching DeepSeek-V3-style bias-based balancing. |
| Multi-token prediction layers | not in its code | not stated | No num_nextn_predict_layers/mtp config key or MTP module exists in the config or code. |
| SwiGLU feed-forward code only | in its code | not stated | Both the dense LagunaMLP and the expert MLP compute act_fn(gate_proj(x)) * up_proj(x) with hidden_act='silu' (SiLU = SwiGLU). show evidence |
token mixer 5
softmax attention 3
“with softplus attention gating and per-layer-type rotary scales”
grouped-query attention 2
“grouped-query attention”
“and softplus-based per-head gating”
hybrid layer stacking 2
“Mixed SWA and global attention layout: 48 layers in a 1:3 global-to-SWA ratio (12 global attention layers, 36 sliding-window layers…”
“+ 48 GA / 64 SWA Q-heads, 𝑘dense=1 (final ablation architecture)”
channel mixer 7
mixture of experts 7
“M.1 and XS.2 are Mixture-of-Experts foundation models.”
“Efficient sparse architecture: Laguna S 2.1 has 118B total parameters and activates 8B parameters per token.”
“we adopted routed expert modulation”
expert routing 2
“a token-choice router with softplus gating over 256 routed experts plus one shared expert”
“LAGUNA XS.2 uses token-choice routing with 8 of 256 experts activated per token”
expert load balancing 1
“For load-balancing, we use an auxiliary loss from Qiu et al. [66] across the non-padding tokens of the global batch.”
positional encoding 2
“with softplus attention gating and per-layer-type rotary scales”
“We employ Rotary Positional Encodings (RoPE) to encode positional information.”
5 more mentioned or evaluated
token mixer “Dense GA, full RoPE, full gating”
positional encoding “+ GA Partial RoPE (50%)”
sliding window attention “+ per-head gating, θswa = 10,000”
sliding window attention “+ SWA-1024 (interleaved 3:1)”
sliding window attention “+ SWA-512”
distillation objective 2
“To mitigate this, we used quantization-aware distillation (QAD) to recover quality; the quantized student was trained to match the traces…”
“We cache final hidden states of the teacher and reconstruct teacher and student logits through the frozen output head.”
optimization 13
optimizer 2
“We use the Muon optimizer [39], specifically the Moonlight variant of Liu et al. [49], across all training stages”
“we adopt Moonlight-style learning-rate scaling [49].”
learning-rate schedule 2
“we adopted a Warmup-Stable-Decay (WSD) learning rate schedule instead of cosine.”
“we therefore fit a WSD-specific scaling law by sweeping four MoE sizes”
training precision 2
“We use BF16 mixed precision training throughout all stages with the master weights in FP32”
“S 2.1 is also our first model where RL was done in FP8 precision, accelerating that part of the training.”
training stability 2
“we apply an exponential moving average (EMA) over the 10 most recent checkpoints”
“we periodically hash model weights across replicas to assert that they remain bit-identical.”
training runtime 5
“an in-flight recovery system detects hangs, NCCL failures, and node losses, replaces affected workers, and resumes from the latest…”
“The weight sync uses NCCL point-to-point primitives over GPUDirect RDMA with no intermediate offload to host memory or remote object…”
“Capacity decisions are made at the granularity of jobs, not nodes.”
“Pre-flight checks stress-test every node before admission”
“the scheduler prefers to respawn it on the same node it was previously bound to”
data sourcing 2
“We shifted from a predominantly high-precision web pipeline toward a substantially higher-recall pipeline designed to preserve diversity…”
“We stream training data rather than reading it from local disk.”
data filtering 9
“For LAGUNA XS.2, the primary rejection path is therefore model-based and deliberately conservative. The pipeline removes documents only…”
“Hence we decompose document quality into a set of independently learnable properties, which we then recombine into a composite…”
“The ranking pipeline therefore leverages model-based annotations to remove high-confidence noise, assigns continuous quality signals, and…”
“For the remaining documents, quality is treated as a ranking signal rather than fully filtered out. Documents are then sorted by a…”
“We densely annotate this region within the N × I space”
“We keep Propella’s native rating ranges, use a PCA-informed subset of its dimensions to de-correlate signals, and then form the composite…”
“The document tags are produced with Propella [36], an open multi-property document annotation model that exposes separate dimensions…”
“we model document quality along two complementary axes: a noise axis (N) ... while an information axis (I) captures whether the document…”
“optionally keeping a single task per repository to enforce codebase diversity”
deduplication 1
“Snapshot-level fuzzy deduplication is preferred because it gives higher performance on knowledge benchmarks.”
synthetic data 10
“where Hᵀ is a dynamic agent-interaction loop over orchestrators, generators, judges [92, 106], and per-step early-exit gates given a…”
“We take a modular approach to synthesis: each pipeline composes from a small set of reusable components among six types”
“We augment agentic tasks with synthetically generated system messages before trajectory generation.”
“We expanded the amount of usable training signal through targeted synthetic rephrasing at scale.”
“A task the teacher cannot solve in one shot becomes a source of biased generation, so we either decompose or strengthen the teacher.”
“We hand our generator chain G anything we already know about the output or how to achieve it through supplying metadata, reducing the…”
“Each Gᵢ transforms the previous output or emits metadata for later stages; the parenthesized V-gate prunes intermediate outputs.”
“We use synthetic data to complement our organic data sources. Synthetic data builds on top of those sources rather than replacing them”
“Software-engineering tasks are mostly grounded in real code history: the largest source reproduces real commits (~38,000 tasks across…”
“Tasks are filtered with a two-sided correctness check that requires the gold solution to pass the tests and an empty solution to fail them”
data mixture & curriculum 5
“AutoMixer operates by training a swarm of proxy models, each trained on a different data composition.”
“To avoid unrealistic shifts toward a small number of dominant sources, we regularize the optimization toward the baseline prior”
“For sampling, we define bucket boundaries by blind pairwise comparisons around candidate thresholds, so documents across a boundary are…”
“Candidate mixtures are sampled as: x ∼ Dirichlet(αx₀) subject to additional constraints: ‖x − x₀‖₁ < ε”
“We drop tasks that were always solved or never solved, and sample the remaining tasks proportional to their(1− trace_rate)”
sequence packing 1
“We pack sequences shorter than the maximum sequence length into a single microbatch.”
tokenization 1
“For initialization, we use subtoken averaging [25, 72] — each new token’s embedding is initialized as the mean of its constituent…”
filed at the root 1
“The majority of what separates S 2.1 from the XS models comes from post-training, in two stages: an SFT stage that bootstraps capabilities…”
supervised fine-tuning 4
“an SFT stage that bootstraps capabilities partly with synthetic data”
“We run for three epochs of 40B tokens each, with early stopping based on evaluation scores.”
“We then run a 100-step warmup phase where the full network except input embeddings and LM head layers is frozen”
“Instruction following: Malibu 2.1 went through additional training to understand and follow user instructions more accurately, especially…”
reinforcement learning algorithm 5
“We train with a token-level REINFORCE surrogate combining importance-sampling-ratio clipping (CISPO [14]) with a length-weighted…”
“The RL prompt distribution spans three families that share a single tool-use API.”
“Code editing: Malibu 2.1 was trained on an expanded set of RLCEF tasks, leading to better downstream performance at generating code edits…”
“for every assistant step whose tool invocation led to the error in tool execution, we apply a small negative reward to exactly the tokens…”
“Moonlight scaling [49] was deactivated during RL for LAGUNA M.1, but used for LAGUNA XS.2.”
reward modelling 4
“For SWE-style tasks this is the repository’s own unit-test suite run against the agent’s patch; for terminal tasks it is the task’s…”
“Rewards are produced by a deterministic chain of checkers applied to every terminated rollout in the following order, with the first…”
“long-horizon credit assignment is carried entirely by the binary verifier at the end of the trajectory, propagated to every token of the…”
“This per-token shaping focuses credit assignment on the failing step and discourages long chains of malformed tool calls before falling…”
rollout & RL infrastructure 8
“We use a token-in, token-out (TITO) API design for the actors used in our RL training. This ensures that token IDs are stable across…”
“Weight updates are configured to block any in-flight rollout step on the update, so no single rollout step straddles a refresh”
“We also cap the staleness of trajectories used for training at 10 optimizer steps, which we empirically found to be a good corner between…”
“Closed-loop interaction; orchestrator decides flow and termination.”
“After every generation step we render the conversation history with this flag enabled and assert that the resulting string exactly matches…”
“More generous rollout budgets: substantially longer timeouts, more tokens per turn, and more turns per task than any model before it…”
“Multi-harness rollouts: the same prompts are rolled out in several agent harnesses, so the model learns behaviors that carry across…”
“Better sandbox infrastructure: RL moved to a new sandboxing service, enabling background processes, selective network blocking to shrink…”
agentic post-training 4
“trained with reinforcement learning inside our agent harness”
“The harness driving RL rollouts is the same one we ship in production: the policy interacts through the same tool API, chat template, and…”
“We intentionally preserve native harness behaviors during data collection — including custom subagent spawning, context compaction…”
“Terminal tasks come from datasets where agents build unseen environments and tasks from a seed.”
mid-training & continual pretraining 1
“The first sub-stage extends the context to 32K tokens; the second extends it to 128K tokens.”
decoding strategy 1
“Speculative decoding: a trained DFlash draft model is available for lower-latency serving”
reasoning control 5
“Native reasoning support: interleaved thinking between tool calls”
“Native reasoning support: You can run requests with thinking off or with max thinking. Max thinking is enabled by default”
“Laguna S 2.1 has two thinking modes: off and max (enabled by default) where it determines the right thinking/test-time compute budget for…”
“We train with persistent thinking history, where all prior reasoning blocks remain visible in the context during each generation step.”
“per-request control via enable_thinking”
KV cache management 1
“Weight broadcasts trigger an inference-side KV-cache reset, so the cache can never mix tokens encoded under different weight versions.”
inference quantization 8
“storing the KV cache in FP8 roughly doubles the number of concurrent trajectories a single replica can carry”
“For INT4 weight quantization (W4A16), we applied AWQ using a calibration set of 128 long-context agentic trajectories. This approach…”
“The KV cache was quantized to FP8 using amax-based scaling with a calibration dataset of 128 long-contexts agentic traces.”
“For FP8 quantization (W8A9), we used a amax-based quantization scheme with dynamic activation scaling that does not require calibration.”
“We adopted a mixed-precision quantization strategy: the first 30 layers were quantized to INT4, while the final 10 layers were quantized…”
“We quantized the model’s MoE layers to FP8, INT4, NVAP, and the KV cache to FP8.”
“We found that using SpinQuant R1 rotation improved the quality of FP8 quantization without working any runtime overhead, and therefore…”
“Poolside worked with NVIDIA to optimize inference from TRT-LLM serving to NVFP4 on Blackwell, down to a single DGX Spark.”
inference scheduling 1
“To manage traffic across Atlas deployments we run an Envoy-based proxy paired with a custom orchestrator that owns session management and…”
context management 1
“keep reasoning_content from prior assistant messages in the message history”
agentic scaffolding 4
“Our parser instead scans only the current turn’s content, so prior reasoning spans in the conversation history do not short-circuit the…”
“In practice, deltas can contain multiple tokens — for example, under speculative decoding or when vLLM’s single-slot output collector…”
“Laguna S 2.1 is guided to use a tool call format where tool calls are marked by XML-like tags:”
“If this occurs, use the checkpointing feature to roll back its changes.”
1 more mentioned or evaluated
agentic scaffolding “Give Laguna M.1 a clear task with the specific outcome you want. Include relevant context such as file paths, error messages, failing…”
Not used 2
reasoning control “We are releasing this model without user-configurable effort (low-medium-high) control today to get Laguna S 2.1 into the hands of users…”
inference quantization “we quantized the inference-audit weights to FP8 with an in-flight block-wise copying: weights were re-quantized on the inference replica…”
filed at the root 1
“maintaining a single code base for research and production ... gives us a crucial advantage”
inference engine 1
“Atlas serves every downstream consumer: automated evaluations, internal inference, customer-facing production traffic, but also crucially…”
training framework 2
“We use Dagster [19] as the central control plane to navigate the directed acyclic graph (DAG) of assets”
“All inputs and configurations to any run ... are expressed as code committed into a single repository.”
infrastructure service 5
“Maintaining a DAG of inputs, runs, and artifacts allows us to surface the complete lineage of any model or artifact in both directions”
“We run a sidecar for HTTP and HTTPS requests to handle these requests instead with a fallback to the default httpbin.org endpoint.”
“We replaced it with an in-house batch scheduler purpose-built around three observations”
“Our internal platform for research and engineering, which we refer to as the Model Factory, allows us to industrialize the model…”
“We vendored (stored copies in internal storage) all external dependencies for tasks fix-ocaml-gc, reshard-c4-data, build-cython-ext…”
evaluation 12
filed at the root 5
“We report mean pass@1 averaged over multiple attempts per task (avg@k), with k per benchmark listed below.”
“Using a mixed set of benchmarks to properly validate that the quantization process does not introduce any unintended side effects…”
“For every benchmark score we publish today, we are releasing full trajectories for every trial in the final evaluation set at…”
“To address this, we run each benchmark four times and report the mean pass@1 across the four runs.”
“we report external model baseline scores only from official release blog posts by the authors or equivalent”
benchmark 1
“Execution-grounded evaluation: benchmarks where each task requires a real software environment that compiles, runs, and tests the model's…”
evaluation harness 3
“using the Laude Institute's Harbor Framework with our agent harness”
“We schedule configurable evaluations automatically for every training experiment, run against in-flight checkpoints which execute on Atlas.”
“evaluations run using pool, Poolside's agent harness, so researchers and developers can reproduce the results on their own hardware”
judge 3
“We developed a dedicated IF judge that decomposes a system message into a list of independent behavioral requirements and scores each…”
“Using an LLM as a Judge (LLMaaJ) calibrated with human labeled trajectories, we tag solutions that potentially reward hack after every eval.”
“we have developed a reward hack judge to detect these specific cheating strategies, which we ran post-hoc-on all candidate evaluation runs…”
other 5
filed at the root 5
“Another core principle of our Model Factory is to minimize the need for repetitive and mechanical manual work as much as possible…”
“Today we use agents to design and run ablation experiments, monitor runs and debug issues, and compile and analyze experiment results.”
“we patched the base task images to remove the git history leak for all affected code_path benchmarks”
“To address this we added a small prompt addendum to the user prompt instructing the model to not use direct solutions found online.”
“Each run is tracked with a unique ID”
unfiled 23
“Read files and folders to understand project structure. Edit files. Run programs, such as linters, compilers, tests, or local services, to…”
“New for S 2.1 is agentic repository installation: given a repository, install every dependency and get the test suite running.”
“We then ran Laguna S 2.1 in an automated research loop that fed each result back to it and pushed it to keep improving.”
“For this task, we instrumented the harness with benchmarks so the model could see exactly where the time and memory went.”
“a successful innovation can be promoted into production by simply flipping a configuration flag.”
“Fixed dependency drift on tasks: build-pmars, crack-7z-hash, mteb-leaderboard, mteb-retrieve, rstan-to-pystan, dna-assembly, dna-insert…”
“For SWE-Bench Pro we made edits focused on fixing verifier selections and the previously mentioned reward hack vulnerabilities.”
“If the known teardown flake occurs with the exact signature seen under the golden patch, and is the only failure, we ignore it.”
“We added in-task retries on components that rely on external services which are known to be flaky”
“We set strict rules: one approach at a time, benchmark after every change, keep only what measurably wins.”
“the weights are available today on Hugging Face under an OpenMDW-1.1 license”
“For each (𝑁, 𝐷 ∈ {60, 120, 240, 480, 960} B) we fit a parabola in log10(lr) space and take the vertex as lr⋆(𝑁, 𝐷)”
“Per-cell fits. Per-(𝑁, lr) power-law fits 𝐿(𝐷) = 𝐿 0 + 𝐴 𝐷 −𝛾 across the five cooldown points are shown in Figure8”
“To fix this we pinned to setuptools^=58.0.0.”
“Poolside team shared three unedited trajectories to show behavior rather than scores.”
“All Laguna S 2.1 agentic benchmarking was completed using our internal fork of the Laude Institute's Harbor Framework with our agent…”
“What we've done in this model is not necessarily add more intelligence, but improve the behaviors that lead to a more capable model: more…”
“Supporting figures and tables for the WSD optimal-LR scaling law described in Section3.1.1.”
“When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it.”
“Poolside team publishes weights in BF16, FP8, INT4, and NVFP4”
3 more mentioned or evaluated
“Malibu Agent can sometimes create test scripts to validate its changes, but forget to delete them. Check to make sure you do not commit…”
“If Malibu Agent is unable to solve an issue, break your request down into smaller, incremental changes, or be more direct. For example…”
“Keep tasks self-contained by planning with the model.”
Further reading
Picked by hand, not extracted: where to read more, not evidence for anything on this page.
- Laguna-S-2.1 architecture card — LLM Architecture Gallery explainer sebastianraschka.com
Sources
The curated source list for this model. "not fetched" means the URL is recorded but its text was not read in the current run.
| Source | Kind | Publisher | Status |
|---|---|---|---|
| poolside.ai/blog/introducing-laguna-s-2-1 | official blog | Poolside | read |
| poolside.ai/assets/laguna/laguna-m1-xs2-technical-report.pdf | technical report | Poolside | read |
| huggingface.co/poolside/Laguna-S-2.1 | model card | Poolside | read |
| docs.poolside.ai/release-notes/models | vendor docs | Poolside | read |
| poolside.ai/models | vendor docs | Poolside | read |
| venturebeat.com/infrastructure/poolside-drops-laguna-s-2-1-an-open-weig… | news | VentureBeat | not fetched |
| marktechpost.com/2026/07/21/poolside-releases-laguna-s-2-1 | news | MarkTechPost | read |
| openrouter.ai/poolside/laguna-s-2.1 | vendor docs | OpenRouter | read |
| finance.yahoo.com/technology/ai/articles/poolside-releases-laguna-2-1-1… | news | Yahoo Finance | read |