Note: Overall leaderboard rankings may not reflect true model quality — individual benchmarks give a clearer picture. ARC-Challenge MMLU GPQA GSM8K Artificial Analysis Intelligence Index v4.1
← Back to leaderboard

BigScience: bloom-560m

8 benchmarks
CompanyBigScience
Country of originFrance
WeightsOpen weights
Parameter count560.0M
Architecturebloom
Download Q8_0 GGUF for llama.cpp
Repository on Hugging Face
Knowledge domainGeneral

Leaderboard position

#1093
Overall rank (of 1121 models)
-641
Elo rating
8
Benchmarks
Top 97.5%
Percentile
#1
In BigScience (of 1 models)
Above: AMD: AMD-135m #1092 Below: Amir: tiny-llm #1094

Benchmark highlights

Best benchmark
WinoGrande 49.9605
Worst benchmark
WikiText-2 (-ppl) -38.0031

Nearest competitor from another company

The model from another company with the most similar benchmark scores.

appeared as nearest in 2 of 8 benchmarks

LiquidAI: LFM2.5-230M
2/8 25.0%
Model Votes % of benchmarks
LiquidAI: LFM2.5-230M 2 25.0%
pleias: baguettotron 1 12.5%
AMD: AMD-135m 1 12.5%
Technology Innovation Institute: Falcon H1 Tiny 90M Instruct 1 12.5%
SupraLabs: Supra-50M BASE 1 12.5%
Qwen: qwen2.5-math-1.5b-instruct 1 12.5%
SupraLabs: Supra-50M-Reasoning 1 12.5%

Benchmarks

WinoGrande 49.9605 ARC-Easy 37.193 HellaSwag 36.25 MMLU 27.6486 TruthfulQA 25.9486 ARC-Challenge 21.7391 -(CoderPPL) -6.9102 WikiText-2 (-ppl) -38.0031

Predicted scores

Estimated from 25 similar models and the global Elo rating. These are projections, not measured results.

Benchmark Predicted Confidence Based on
PIQA ~62.01 65% Google: Gemma 3 270M Instruct, raincandy-u: rain-v2, AMD: AMD-135m
IFEval ~61.73 63% Google: Gemma 3 270M Instruct, LiquidAI: LFM2.5-230M, LiquidAI: lfm-2.5-350m
GSM8K ~24.37 63% Google: Gemma 3 270M Instruct, LiquidAI: lfm-2.5-350m, Hugging Face Smol Models Research: smollm2-135m-instruct
CowQA ~20.58 62% LiquidAI: LFM2.5-230M, Sapbot: Grok 4 Distill SmolLM2 135M, NON AI - НЕ ИИ: Random baseline - Случайный счёт
GPQA Diamond ~26.34 61% LiquidAI: LFM2.5-230M, LiquidAI: lfm-2.5-350m, NON AI - НЕ ИИ: Random baseline - Случайный счёт
IFBench ~30.49 61% LiquidAI: LFM2.5-230M, LiquidAI: lfm-2.5-350m, OpenBMB: minicpm5-1b
Artificial Analysis Intelligence Index v4.1 (Maximum Reasoning) ~1 61% Google: Gemma 3 270M Instruct, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 0.8B
Artificial Analysis Intelligence Index v4.0 (Maximum Reasoning) ~4.7 61% Google: Gemma 3 270M Instruct, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 0.8B
MMLU-Pro ~26.29 60% LiquidAI: lfm-2.5-350m, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 0.8B
Terminal-Bench v2.1 ~8.6 60% Xiaomi: MiMo-V2.5 Pro
SWE-bench Pro ~5.24 60% Xiaomi: MiMo-V2.5 Pro
SciCode ~3.9 60% Xiaomi: MiMo-V2.5 Pro
AA-LCR ~13.88 60% Xiaomi: MiMo-V2.5 Pro
Artificial Analysis Intelligence Index v4.1.1 (Maximum Reasoning) ~3 60% Xiaomi: MiMo-V2.5 Pro
Text Arena (Snapshot from 20 june 2026) ~1151.27 60% Hugging Face Smol Models Research: smollm2-1.7b-instruct, Xiaomi: MiMo-V2.5 Pro
BBH ~34.32 59% Hugging Face Smol Models Research: smollm2-135m-instruct, OpenBMB: minicpm5-1b, Qwen: qwen2.5-0.5b
Artificial Analysis Intelligence Index v4.3.2 ~5 59% Google: Gemma 3 270M Instruct, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 0.8B
SuperGPQA ~7.52 59% OpenBMB: minicpm5-1b, Qwen: Qwen3.5 0.8B, Qwen: qwen2.5-0.5b
Claw-Eval ~12.78 58% Xiaomi: MiMo-V2.5 Pro
BrowseComp ~0.82 58% NON AI - НЕ ИИ: Random baseline - Случайный счёт, NON AI - НЕ ИИ: Human baseline - Человеческий результат
GPQA ~10.6 58% Qwen: Qwen3.5 0.8B, Qwen: qwen2.5-0.5b, Google: gemma-3-1b-pt
Terminal-Bench v2.0 ~0 58% NON AI - НЕ ИИ: Random baseline - Случайный счёт, Xiaomi: MiMo-V2.5 Pro, NON AI - НЕ ИИ: Human baseline - Человеческий результат
MBPP ~25.33 57% Qwen: qwen2.5-0.5b, Google: gemma-3-1b-pt, Qwen: qwen3-0.6b-base
MATH ~3.66 57% Qwen: qwen2.5-0.5b, Google: gemma-3-1b-pt, Qwen: qwen3-0.6b-base
MATH-500 ~66.06 56% OpenBMB: minicpm5-1b, LiquidAI: lfm-2.5-8b-a1b, Hugging Face Smol Models Research: smollm3-3b