[whatmodel.sh](../../README.md) / OpenChat 3.5 # OpenChat 3.5 OpenChat · released 2023-10 · 7B params · apache-2.0 **score 43.3** ±1.4 · #511 of 732 · from 7 benchmark results across 4 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at _{● marks OpenChat 3.5 among every model on that task or test, worst to best.}

task            vs the others          rank  counts  tests
coding          ▁▁▁▃▅●█▇▅▄▃▁   #349 of 469*     16%      1
chat & writing  ▁▁▁▃●▄▆█▆▇▇▃   #328 of 402*     39%      2
math            █··········●       #1 of 2*     13%      1

coding · #349 of 469 · 16% of its score

Beats [LFM2.5 2.6B](lfm2-5-2-6b.md), [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md), [Qwen3-VL 2B Thinking](qwen3-vl-2b-thinking.md) here, though they rank higher overall. Trails [Phi-3-mini-128k-instruct (Apr 2024)](phi-3-mini-128k-instruct.md), [DeepSeek LLM 67B Chat](deepseek-llm-67b-chat.md) here, though they rank lower overall.

test               result  vs the others in its run    difficulty      counts  why
HumanEval (maker)   55.5%  ▁▁▁▁▂▂▆●█▅▂▁ beats 106/205  easy · avg 72%     16%  sharp · at its level · small run

chat & writing · #328 of 402 · 39% of its score

Beats [Gemma 4 E4B](gemma-4-e4b.md), [OpenThinker3 7B](openthinker3-7b.md), [Code Llama 70B Instruct](codellama-70b-instruct.md) here, though they rank higher overall. Trails [LFM2 350M](lfm2-350m.md), [Gemma 3 1B](gemma-3-1b.md), [Qwen3 0.6B](qwen3-0-6b.md) here, though they rank lower overall.

test                                result  vs the others in its run   difficulty  counts  why
LMArena (text)                      1183.1  ▁▁▁▂●▃▃▆▆▇█▃ beats 59/385  avg 1260       20%  blunt
Polyrating rating (bias-aware re-…  1101.0  ▂▁▂▁▃▇●▆▇▄▂▂ beats 47/95   avg 1187       19%  typical

math · #1 of 2 · 13% of its score

test           result  vs the others in its run    difficulty           counts  why
GSM8K (maker)   77.3%  ▂▂▂▂▃▄▅▇█●▆▄ beats 134/206  very easy · avg 88%     13%  sharp · at its level · small run

other tests (3)

test                                result  vs the others in its run    difficulty        counts  why
BIG-Bench Hard (maker)               63.5%  ▁▁▁▁▂▄▅▆●▃▃▁ beats 93/155   easy · avg 76%       12%  sharp · at its level · small run
MMLU (maker)                         64.3%  ▁▁▁▁▁▂▂▅●█▆▄ beats 101/214  easy · avg 75%       10%  at its level · small run
EQ-Bench v1 (60 questions, paper …   37.1%  ▂··▄·▂●▄▂█·▂ beats 5/14     medium · avg 53%    8.7%  at its level

## 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
OpenChat 3.5 model card (submitted / ma…        4   52% ▇▇▇▇▇▇
LMArena leaderboard dataset                     1   20% ▇▇
Polyrating: A Cost-Effective and Bias-A…        1   19% ▇▇
EQ-Bench: An Emotional Intelligence Ben…        1    9% ▇
prior: pull toward the middle                        1% ▇
weighted average                                   = 43.3

_{Licences: OpenChat 3.5 model card ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · EQ-Bench: An Emotional Intelligence Benchmark for Large Language Models ([CC BY 4.0](../../licenses/paper-2312.06281v2-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 OpenChat 3.5 (Q4_K_M, needs ~4.4 GB, from [huggingface.co/TheBloke](https://huggingface.co/TheBloke/openchat_3.5-GGUF/tree/9145d8078ca54b163b4b7f7d45cf525b63ee77f8)) ```sh llama cli -hf TheBloke/openchat_3.5-GGUF:Q4_K_M ``` 3. Or start a local server and open the address it prints (http://localhost:9931) ```sh llama serve -hf TheBloke/openchat_3.5-GGUF:Q4_K_M ``` Other versions: Q8_0 (~7.7 GB) `llama cli -hf TheBloke/openchat_3.5-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 OpenChat 3.5 (Q4_K_M, needs ~4.4 GB, from [huggingface.co/TheBloke](https://huggingface.co/TheBloke/openchat_3.5-GGUF/tree/9145d8078ca54b163b4b7f7d45cf525b63ee77f8)) ```sh lms get https://huggingface.co/TheBloke/openchat_3.5-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 (~7.7 GB) `lms get https://huggingface.co/TheBloke/openchat_3.5-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 OpenChat 3.5 (Q4_K_M, needs ~4.7 GB, from [ollama.com](https://ollama.com/library/openchat/tags)) ```sh ollama run openchat:7b-v3.5-q4_K_M ``` Other versions: Q8_0 (~8.9 GB) `ollama run openchat:7b-v3.5-q8_0` Model repository and documentation: [huggingface.co/openchat](https://huggingface.co/openchat/openchat_3.5) [interactive version](https://whatmodel.sh/model/openchat-3-5) · [JSON](https://whatmodel.sh/api?format=json&all=1)