Lara Mini Base Report (base-model protocol) - MATH (4-shot): leaderboard

Harness: lm-evaluation-harness 0.4.14.dev0; task configs unchanged except shots (A4.T17); no chat template; identical prompts/settings/answer extraction for every model (§5.1; Appendix D). Backends: Log-likelihood tasks: HF Transformers backend, fp32 log-softmax. Generative tasks: vLLM, bf16, greedy decoding, 4,096-token context; backends agree within 0.6 points on checked tasks (§5.1; Appendix D). Shots and metrics: Per A4.T17 (e.g. HellaSwag 10-shot acc_norm, MMLU 5-shot, MMLU-Pro 5-shot CoT exact match, GSM8K 5-shot EM, MATH 4-shot Math-Verify, MGSM 8-shot CoT over 11 languages, MBPP/MBPP+ 3-shot pass@1, CRUXEval 2-shot pass@1) (A4.T17). Answer extraction: Stricter parsers for GSM8K, MGSM (native CoT answer phrase) and MBPP+ (must pass MBPP and EvalPlus tests); MATH Minerva prompts with Math-Verify over 5,000 problems; CRUXEval direct prediction with official parser (Appendix D Answer extraction). Evaluated Lara checkpoint: Released checkpoint after the 100B-token long-context stage (translated/Lara-4-Mini-Base @ c998c0e1251a734eeff364d532330790a4f93c65 as captured; revision used by the authors not stated) (§5 introduction; §7 Release). Long context: RULER (13 tasks), ONERULER (4 retrieval tasks, 26 languages), HELMET (4 categories) at 32k and 128k with base-model modifications in A4.T24 (§5.7; Appendix D.3; A4.T24). Runs / seeds: not stated; scoring is log-likelihood or greedy generation (not stated). Benchmark availability: T-HellaSwag is the authors' translation, announced as "will release" at v1 (§5.2; §7). This board: Full benchmark (MATH: 5,000 problems; others not printed); Single greedy generation per item (vLLM bf16, 4,096-token context); run count not stated. lm-evaluation-harness 0.4.14.dev0, 4-shot (A4.T17); answer-extraction overrides in Appendix D. Metric: Math-Verify equivalence, Minerva prompts, 5,000 problems over seven subjects (%) Higher is better.

Metric: Math-Verify equivalence, Minerva prompts, 5,000 problems over seven subjects (%). Source: arxiv.org. Saturation forecast: Around February 2027. 55 models tracked. Updated 14 Nov 2026.

Top models

#ModelScoreOverall rank
1Hunyuan-4B-Pretrain74.8#2455
2Nemotron-Flash-3B67.7#6435
3Lara-4-Mini-Base60.9#5152
4Qwen3-8B-Base55.3#1774
5Qwen3.5-4B-Base54.2#2478
6limite-1B-base-soup53.8
7limite-1B-base53.4#8819
8Qwen3-4B-Base52.8#2469
9Qwen 2.5 7B49.9#2829
10Falcon-H1-1.5B-Deep-Base49.2#5677
11Falcon-H1-1.5B-Base46.6#6880
12LFM2.5-8B-A1B-Base45.8#6243
13granite-4.0-h-tiny-base45.4#5994
14Nemotron-Flash-1B44.9#8677
15LFM3-1.7B-Base43.6#6594

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=lara-mini-base-report-base-model-protocol-math-4-shot · How It Works · Data refreshed daily, snapshot 2026-11-16.