# Changelog

## Score

- CAI 50 → 41 (-9.0)
- Rubric changed (rubric-2026.10.1 → rubric-2026.10.5) — scores are not directly comparable.

## Lenses

- Code Health 67 → 75 (+8.1)
- Architecture 100 → 100 (+0.0)
- Maturity 52 → 31 (-21.0)
- Readiness 32 → 33 (+0.7)
- Security 81 → 44 (-36.6)
- Performance 100 → 100 (+0.0)

## Resolved (1)

- Banned license: mdtex2html

## New (21)

- Duplicate batch processing methods with inconsistent signatures. `response_batch` and `chat_batch` perform the same batched generation but differ in the presence of the `tokenizer` argument, mirroring the inconsistency in the non-batched methods.
- Inconsistent naming and parameter handling for generation methods. `response` and `chat` perform similar generation tasks but differ in whether they require an explicit `tokenizer` argument and how they handle prompt formatting. Similarly, `stream_response` and `stream_chat` duplicate this pattern. The distinction between 'response' and 'chat' is unclear (likely internal vs external prompt handling), creating a confusing API surface.
- Orphaned file (cli_demo.py)
- Orphaned file (dcu-support/package/fastllm_pytools/hf_model.py)
- Orphaned file (dcu-support/package/fastllm_pytools/torch2flm.py)
- Orphaned file (eval/evaluate_ceval.py)
- Orphaned file (eval/evaluate_chat_ceval.py)
- Orphaned file (eval/evaluate_chat_gsm8k.py)
- Orphaned file (eval/evaluate_chat_humaneval.py)
- Orphaned file (eval/evaluate_chat_mmlu.py)
- Orphaned file (eval/evaluate_cmmlu.py)
- Orphaned file (eval/evaluate_gsm8k.py)
- Orphaned file (eval/evaluate_humaneval.py)
- Orphaned file (eval/evaluate_mmlu.py)
- Orphaned file (eval/evaluate_plugin.py)
- Orphaned file (finetune.py)
- Orphaned file (recipes/inference/vllm/vllm_wrapper.py)
- Orphaned file (run_gptq.py)
- Orphaned file (web_demo.py)
- Orphaned knowledge (openai_api.py)
- …and 1 more