[whatmodel.sh](../../README.md) / Falcon 40B Instruct # Falcon 40B Instruct TII · released 2023-05 · 41.8B params · apache-2.0 **score 31.2** ±2.0 · #654 of 732 · from 7 benchmark results across 6 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at ● marks Falcon 40B Instruct among every model on that task or test, worst to best. 
task            vs the others         rank  counts  tests
chat & writing  ▁▁●▃▆▄▆█▆▇▇▃   #393 of 402     85%      3
chat & writing · #393 of 402 · 85% of its score Beats [Phi-2 (2.7B)](phi-2.md) here, though they rank higher overall. Trails [Gemma 3 270M IT](gemma-3-270m.md), [SmolLM2 360M Instruct](smollm2-360m.md), [Granite 4.0 350M](granite-4-0-350m.md) here, though they rank lower overall. 
test                                                      result  vs the others in its run   difficulty                counts  why
AlpacaEval 2 length-controlled                              5.6%  ▁▂●▃▆▆█▆▄▄▁▁ beats 10/122  medium · avg 35%             42%  sharp · too hard for it
MT-Bench (GPT-4 single-answer gra…  5.1700 MT-Bench score (1-10)  ▆▂▂▆●▄██▂▄▄▆ beats 9/28    avg 8.539 MT-Bench scor…     31%  sharp · noisy
AlpacaEval 1.0 win rate (vs text-…                         45.7%  ▂▂·●·▇▃▆█▆▄▂ beats 5/45    very easy · avg 94%          12%  sharp · at its level · noisy
other tests (4) 
test                                result  vs the others in its run  difficulty           counts  why
clembench v1.0 clemscore (self-pl…    0.7%  ●······▄·█·▄ beats 0/4    medium · avg 38%       7.2%  sharp · too hard for it
This is not a Dataset (negation),…   54.7%  █▁····▃●·▁·▁ beats 12/14  hard · avg 66%         4.2%  sharp · too hard for it · noisy
TruthfulQA MC2                       52.5%  ▅▃█▆▁●▅···▁▁ beats 24/31  medium · avg 57%       1.2%  blunt · at its level
FANToM BeliefQ [Choice], short co…   54.3%  █·····●····▁ beats 10/11  very hard · avg 55%    0.5%  sharp · too hard for it · noisy
## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its *counts* column above; here it is added up by who ran the tests. [Method](../method.md) 
source                                    results  share of the score
AlpacaEval 1.0 and 2.0 leaderboards             2   54% ▇▇▇▇▇▇▇
LMSYS Chatbot Arena leaderboard Space: …        1   31% ▇▇▇▇
clembench: Using Game Play to Evaluate …        1    7% ▇
This is not a Dataset: A Large Negation…        1    4% ▇
Hallucinations Leaderboard results              1    1% ▇
FANToM: A Benchmark for Stress-testing …        1    1% ▇
prior: pull toward the middle                        2% ▇
weighted average                                   = 31.2
Licences: AlpacaEval 1.0 and 2.0 leaderboards ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)) · LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)) · clembench: Using Game Play to Evaluate Chat-Optimized Language Models as Conversational Agents ([CC BY 4.0](../../licenses/acl-2023-emnlp-main-689-CC-BY-4-0.txt)) · This is not a Dataset: A Large Negation Benchmark to Challenge Large Language Models ([CC BY 4.0](../../licenses/acl-2023-emnlp-main-531-CC-BY-4-0.txt)) · Hallucinations Leaderboard results ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)) · FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions ([CC BY 4.0](../../licenses/acl-2023-emnlp-main-890-CC-BY-4-0.txt)) ## How to run it ### llama.cpp (recommended) 1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md) ```sh # macOS / Linux curl -LsSf https://llama.app/install.sh | sh # Windows irm https://llama.app/install.ps1 | iex ``` 2. Run Falcon 40B Instruct (Q4_K_M, needs ~26 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/falcon-40b-instruct-GGUF/tree/d1c134a32112db8239fd4a5ebe772d0931b23067)) ```sh llama cli -hf mradermacher/falcon-40b-instruct-GGUF:Q4_K_M ``` 3. Or start a local server and open the address it prints (http://localhost:9931) ```sh llama serve -hf mradermacher/falcon-40b-instruct-GGUF:Q4_K_M ``` Other versions: Q8_0 (~45 GB) `llama cli -hf mradermacher/falcon-40b-instruct-GGUF:Q8_0` ### LM Studio 1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download) 2. Run Falcon 40B Instruct (Q4_K_M, needs ~26 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/falcon-40b-instruct-GGUF/tree/d1c134a32112db8239fd4a5ebe772d0931b23067)) ```sh lms get https://huggingface.co/mradermacher/falcon-40b-instruct-GGUF@Q4_K_M -y ``` 3. Then run lms ls to see its name, and lms chat followed by that name Other versions: Q8_0 (~45 GB) `lms get https://huggingface.co/mradermacher/falcon-40b-instruct-GGUF@Q8_0 -y` ### Ollama 1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download) ```sh # Linux curl -fsSL https://ollama.com/install.sh | sh ``` 2. Run Falcon 40B Instruct (Q4_0, needs ~24 GB, from [ollama.com](https://ollama.com/library/falcon/tags)) ```sh ollama run falcon:40b ``` Model repository and documentation: [huggingface.co/tiiuae](https://huggingface.co/tiiuae/falcon-40b-instruct) [interactive version](https://whatmodel.sh/model/falcon-40b-instruct) · [JSON](https://whatmodel.sh/api?format=json&all=1)