smolcua - Computer-Use Agents

smolcua evaluates text-based computer-use reasoning: tool selection, next actions, strategy switching, instruction following, ambiguity, coding, structured output, coordinates, multi-constraint formatting, and state tracking. It is a synthetic text benchmark, not live UI automation or a visual computer-use environment.

1,000 questions per pass, five passes, 5,000 total. Each question contains four options and an explicit response contract. Models generate the answer; there is no forced-choice logit shortcut, output repair, or answer extraction.

Current edition

The default is corrected v1.1, recorded as edition: local-corrected in its manifest for continuity with the local audit. It repairs 250 incomplete or ambiguous prompts across five templates. Original v1.0 data remains available in runs/v1_seed_20261023 and the original_v1_0 dataset configuration. Options, answer labels, and weights are unchanged. The editions have different hashes; do not report corrected scores as original v1.0 scores.

See LOCAL_CHANGES.md and the per-question corrections. Restored specifications were reconstructed during dataset authoring from the original authored answer keys. Inference never uses those keys to construct prompts or select answers.

Run

Download this repository's toolkit files, then:

pip install -r requirements-transformers.txt
python run_smolcua.py --model YOUR_ORG/YOUR_MODEL --backend transformers \
  --dataset dotlabs/smolcua --prompt-mode plain --max-model-len 4096 \
  --output-dir results/my-model --score

plain is appropriate for base models and provides zero-shot prompts. Use --prompt-mode auto or chat for instruction-tuned models with chat templates. There are no few-shot exemplars in either mode. --dataset-revision COMMIT pins the dataset; --revision COMMIT pins model/tokenizer loading.

The repository is a public self-scoring edition, including files under private/ for evaluation. The directory name does not make them confidential. It is not a blind leaderboard. Disclose exposure to answer keys when reporting results; do not use test questions/keys for training.

Pretokenization and inference

• All selected prompts are compiled in CPU batches before Transformers model weights are loaded. The cache key includes actual prompt contents, dataset identity, tokenizer vocabulary/configuration, and prompt formatting.
• Cache hits reuse exact input IDs. No prompt truncation is allowed.
• Transformers uses length/budget buckets, greedy cached decoding, a default batch size of 32, and CUDA OOM backoff down to one. Supported models compute only the last-position logits during generation.
• Tokenization, model loading, prefill, decode tokens, completed questions, elapsed time, throughput, and ETA appear in flushed progress logs. Long calls get heartbeat updates; no fabricated completion percentage is shown.
• EOS counts as a generated token; padding after a completed sequence does not. Raw response text is not printed in progress logs.
• vLLM supports pretokenized inputs and its native generation progress. API mode uses concurrent requests and server-side tokenization.

# Compile prompts without allocating model weights:
python run_smolcua.py --model YOUR_MODEL --pretokenize-only --prompt-mode plain

# Larger batches; CUDA OOM backoff preserves completed answers:
python run_smolcua.py --model YOUR_MODEL --backend transformers \
  --batch-size 64 --progress-interval 5 --output-dir results/model --score

Performance depends on model, hardware, padding, and output lengths. Batched floating-point inference can differ from single-item inference near logit ties; record the batch size and precision. This release does not claim a measured GPU speedup. See RUNNING_AND_PUBLISHING.md for settings.

Scores and outputs

• Mean weighted score (0–100), with category-specific weights.
• Worst-pass exact-answer reliability: 100 * min(correct per pass) / 1000.
• Per-pass, category, response-format, and consistency breakdowns.
• Raw answers, dataset hash/edition, configuration, token-cache statistics, actual token counts, generation batching statistics, and throughput.

Output files are predictions.jsonl, run_metadata.json, performance.json, and score.json. --score requires all 5,000 responses; partial requests are rejected before model loading. The standalone scorer can assess missing answers as incorrect. Strict JSON contracts reject duplicate keys.

Weights are research-informed and have not been calibrated against a large representative model sample. v1.1 retains v1.0 weights despite clearer wording. Reference-answer and empty-answer tests validate scorer behavior; they are not model performance results.