--- title: "lm-evaluation-harness vs jiwer" type: "comparison" canonical_url: "https://www.graphcanon.com/compare/eleutherai-lm-evaluation-harness-vs-jitsi-jiwer" tools: ["eleutherai-lm-evaluation-harness", "jitsi-jiwer"] --- # lm-evaluation-harness vs jiwer *GraphCanon updated Aug 7, 2026* ## Verdict Pick lm-evaluation-harness if lm-evaluation-harness is a Python framework for evaluating language models in various parallelism modes using different checkpoint formats, compatible with the Megatron-LM backend; pick jiwer if a Python library for assessing speech-to-text systems with word error rate measures. [lm-evaluation-harness](https://www.eleuther.ai) reports 14k GitHub stars, 3.5k forks, and 938 open issues, last pushed Jul 13, 2026. [jiwer](https://github.com/jitsi/jiwer) has 917 stars, 107 forks, and 19 open issues, last pushed Apr 16, 2026. Figures are from public GitHub metadata via [lm-evaluation-harness's repository](https://github.com/EleutherAI/lm-evaluation-harness) and [jiwer's repository](https://github.com/jitsi/jiwer). | | [lm-evaluation-harness](/tools/eleutherai-lm-evaluation-harness.md) | [jiwer](/tools/jitsi-jiwer.md) | | --- | --- | --- | | Tagline | A framework for few-shot evaluation of language models. | Evaluate speech-to-text systems with word error rate (WER) measures | | Stars | 13,560 | 917 | | Forks | 3,467 | 107 | | Open issues | 938 | 19 | | Language | Python | Python | | Adopt for | lm-evaluation-harness is a Python framework for evaluating language models in various parallelism modes using different checkpoint formats, compatible with the Megatron-LM backend. | A Python library for assessing speech-to-text systems with word error rate measures. | | Persona | - | - | | Runtime | - | - | | License | MIT | Apache-2.0 | | Categories | Evaluation & Observability | Evaluation & Observability | ## Trust and health _Sourced signals - not a safety guarantee. No winner column._ | | [lm-evaluation-harness](/tools/eleutherai-lm-evaluation-harness.md) | [jiwer](/tools/jitsi-jiwer.md) | | --- | --- | --- | | Maintenance | Active (82%) | Slowing (36%) | | Days since push | 24d | 105d | | Open issues (now) | 938 | 19 | | Full report | [trust report](/tools/eleutherai-lm-evaluation-harness/trust.md) | [trust report](/tools/jitsi-jiwer/trust.md) | ## Shared compatibility - **Python**: [lm-evaluation-harness](/tools/eleutherai-lm-evaluation-harness.md) - Python runtime; [jiwer](/tools/jitsi-jiwer.md) - Python runtime ## Decision facts: lm-evaluation-harness - **Adopt for:** lm-evaluation-harness is a Python framework for evaluating language models in various parallelism modes using different checkpoint formats, compatible with the Megatron-LM backend. ## Decision facts: jiwer - **Adopt for:** A Python library for assessing speech-to-text systems with word error rate measures. ## Choose when ### Choose lm-evaluation-harness if… - License: lm-evaluation-harness is MIT, jiwer is Apache-2.0. - Tags unique to lm-evaluation-harness: data-parallelism, evaluation-framework, expert-parallelism, language-model. - - When you need to evaluate large language models across multiple GPUs in data or tensor parallel configurations. ### Choose jiwer if… - License: jiwer is Apache-2.0, lm-evaluation-harness is MIT. - Tags unique to jiwer: automatic-speech-recognition, evaluation-metrics, python3, speech-to-text. - For accurate speech recognition evaluation needing direct WER calculations ## When NOT to use lm-evaluation-harness - - If your evaluation setup requires pipeline parallelism not currently supported by this framework. ## When NOT to use jiwer - Avoid if your project requires only basic, non-word level performance indicators - Not suitable for evaluating text-to-speech systems, as it focuses exclusively on speech-to-text output assessment ## Common questions ### What is the difference between lm-evaluation-harness and jiwer? lm-evaluation-harness: A framework for few-shot evaluation of language models.. jiwer: Evaluate speech-to-text systems with word error rate (WER) measures. See the comparison table for live GitHub stats and shared categories. ### When should I choose lm-evaluation-harness over jiwer? Choose lm-evaluation-harness over jiwer when License: lm-evaluation-harness is MIT, jiwer is Apache-2.0; Tags unique to lm-evaluation-harness: data-parallelism, evaluation-framework, expert-parallelism, language-model; - When you need to evaluate large language models across multiple GPUs in data or tensor parallel configurations. ### When should I choose jiwer over lm-evaluation-harness? Choose jiwer over lm-evaluation-harness when License: jiwer is Apache-2.0, lm-evaluation-harness is MIT; Tags unique to jiwer: automatic-speech-recognition, evaluation-metrics, python3, speech-to-text; For accurate speech recognition evaluation needing direct WER calculations. ### When should I avoid lm-evaluation-harness? - If your evaluation setup requires pipeline parallelism not currently supported by this framework. ### When should I avoid jiwer? Avoid if your project requires only basic, non-word level performance indicators Not suitable for evaluating text-to-speech systems, as it focuses exclusively on speech-to-text output assessment ### Is lm-evaluation-harness or jiwer more popular on GitHub? lm-evaluation-harness has more GitHub stars (13,560 vs 917). Stars measure visibility, not whether either tool fits your constraints. ### Are lm-evaluation-harness and jiwer open source? Yes - both are open-source projects on GitHub (lm-evaluation-harness: MIT, jiwer: Apache-2.0). ### Where can I find alternatives to lm-evaluation-harness or jiwer? GraphCanon lists graph-backed alternatives at [lm-evaluation-harness alternatives](/tools/eleutherai-lm-evaluation-harness/alternatives) and [jiwer alternatives](/tools/jitsi-jiwer/alternatives) ([lm-evaluation-harness markdown twin](/tools/eleutherai-lm-evaluation-harness/alternatives.md), [jiwer markdown twin](/tools/jitsi-jiwer/alternatives.md)), ranked by typed relationship edges rather than popularity votes. ### Is there a machine-readable version of this comparison? Yes. The markdown twin at [this comparison](/compare/eleutherai-lm-evaluation-harness-vs-jitsi-jiwer.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers. ### Which is better maintained, lm-evaluation-harness or jiwer? lm-evaluation-harness: Active. jiwer: Slowing. Compare maintenance labels, days since push, and release cadence in the trust section below - stars alone do not measure maintenance. ### Where are the full trust reports for lm-evaluation-harness and jiwer? GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [lm-evaluation-harness trust report](/tools/eleutherai-lm-evaluation-harness/trust); [jiwer trust report](/tools/jitsi-jiwer/trust). --- **Machine-readable endpoints** - JSON: [`/api/graphcanon/graph?tool=eleutherai-lm-evaluation-harness`](/api/graphcanon/graph?tool=eleutherai-lm-evaluation-harness) - LLM index: [/llms.txt](/llms.txt) - Full corpus: [/llms-full.txt](/llms-full.txt) _GraphCanon - The knowledge graph for AI development. https://www.graphcanon.com/_