whatmodel.sh / Yi 34B Chat

Yi 34B Chat

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Yi 34B Chat is an open-weights language model from 01.AI, released November 2023. Its published size is 34.39B parameters and a 4k-token context window. Licence: apache-2.0.

Catalogue identity notes

Comparison model from the Epoch AI internal-evaluation CSVs (Epoch release date 2023-11-22). Original Yi-34B-Chat (November 2023); distinct from Yi-1.5-34B-Chat. 34,388,917,248 parameters. Note ollama yi:34b-chat now points to v1.5; the non-v1.5 quant tags are the original (their licence layer differs from the v1.5 tags).

43.2±1.1#479 of 719 LLMsfrom 8 benchmark results across 7 sources

$ how it compares

The LLMs ranked either side of Yi 34B Chat, out of 828 with a score, then more from 01.AI.

#modelscorewhereRAMreleased
566Phi-3.5-mini-instruct Microsoft43±1local~3.7GB2024-08
567Hermes 2 Theta Llama 3 8B Nous Research43±2local~6.6GB2024-05
568Yi 34B Chat 01.AI43±1local~23GB2023-11
569CogVLM2 Llama3 Chat 19B Zhipu AI (THUDM)43±3local~43GB2024-05
570Code Llama 70B (base) Meta43±1local~28GB2024-01
more from 01.AI
308Yi-Lightning 01.AI55±1api price unknown2024-10
373Yi-Large Preview 01.AI52±1comparison only2024-05
396Yi-Large 01.AI51±1comparison only2024-05

see all LLMs · find the ones that fit your machine

$ what it's good at

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #312 of 393 · 81% of its score · 5 tests

Beats Gemma 4 E4B, OpenThinker3 7B, Gemma 4 E2B here, though they rank higher overall. Trails LFM2 350M, Gemma 3 1B, Flan-T5 XXL here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
AlpacaEval 2 length-controlled27.2% ✎▁▂▂▃▆▆█▆▄▄▁▁ beats 72/122medium · avg 39%▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 34%typical
Arena-Hard (v0.1, GPT-4-1106 judge)23.1% ✎▂▁▁▅█▅▄▃▃▅▃▁ beats 19/60medium · avg 52%▇▇▇▇▇▇▇▇▇ 23%sharp · at its level
Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024)1101.0 ✎▂▁▂▁▃▇█▆▇▄▂▂ beats 46/95avg 1201▇▇▇▇ 11%blunt
LMArena (text)1183.7 ✎▁▁▁▂▃▃▃▆▆▇█▃ beats 61/379avg 1268▇▇▇▇ 11%blunt
AlpacaEval 1.0 win rate (vs text-davinci-003, GPT-4 judge)94.1% ✎▂▂·▆·▇▃▆█▆▄▂ beats 39/45very easy · avg 96%▇ 2.3%sharp · noisy
math▁▂▃▅▆▅▄█▄▁▁▁ #370 of 398 * · 11% of its score · 1 test

Trails SmolLM2 1.7B Instruct, Gemma 3 1B, OLMo 2 1B Instruct (0425) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
MATH level 55.1% ✎▃▂▅▆▄█▆▃▅▂▃▄ beats 2/92hard · avg 32%▇▇▇▇ 11%sharp
other tests3 not in a task · 7.4% of its score
testresultvs the others in its rundifficultycountswhy
EQ-Bench v1 (60 questions, paper run)51.0% ✎▂··▄·▂▆▄▂█·▂ beats 11/14medium · avg 57%▇▇ 4.9%at its level
Open PT-LLM Leaderboard average (9 Portuguese tasks)70.8% ✎▃▁▁▁▁▁▁▁▂█▅▁ beats 173/276medium · avg 73%▇ 2.4%at its level · noisy
GPQA Diamond14.7% ✎▁▁▁▁▂▄▅▅███▄ beats 0/310hard · avg 40%—failed run: Well below the 25% guess rate. Peers of similar age and size score 0.25 to 0.27 in the same Epoch run (deepseek-llm-67b-chat 0.246, Llama-2-70b-chat 0.263), which suggests answer-format failures.

$ where to run it

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run Yi 34B Chat · needs ~20.7 GB · from huggingface.co/TheBloke
    chat in the terminal
    llama cli -hf TheBloke/Yi-34B-Chat-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf TheBloke/Yi-34B-Chat-GGUF:Q4_K_M

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run Yi 34B Chat · needs ~20.7 GB · from huggingface.co/TheBloke

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/TheBloke/Yi-34B-Chat-GGUF@Q4_K_M -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run Yi 34B Chat · needs ~21.13 GB · from ollama.com
    chat in the terminal
    ollama run yi:34b-chat-q4_K_M

Official weights and documentation: huggingface.co/01-ai ↗

$ how the score was worked out · AlpacaEval 1.0 and 2.0 leader… 37%, Arena-Hard-Auto v0.1 leaderbo… 23%, Polyrating: A Cost-Effective… 11% and 4 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The 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's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)2▇▇▇▇▇▇▇▇▇▇▇ 37%Apache-2.0
Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)1▇▇▇▇▇▇▇ 23%Apache-2.0
Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)1▇▇▇ 11%CC-BY-4.0
Epoch AI — internal evaluations1▇▇▇ 11%CC BY 4.0
LMArena leaderboard dataset1▇▇▇ 11%CC BY 4.0
EQ-Bench: An Emotional Intelligence Benchmark for Large Language Models — authors' own runs1▇ 5%CC BY 4.0
Open Portuguese LLM Leaderboard (archived)1▇ 2%Apache-2.0
prior · pull toward the middle▇ 1%
weighted average= 43.2

not counted (1)

Epoch AI — internal evaluationsGPQA Diamond [correct]14.7%failed run: Well below the 25% guess rate. Peers of similar age and size score 0.25 to 0.27 in the same Epoch run (deepseek-llm-67b-chat 0.246, Llama-2-70b-chat 0.263), which suggests answer-format failures.

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json