affordance

Showcase  /  Compatibility test

740 models' prompt formats rebuilt in Affordance, each matching the official one token for token

Every chat model was trained on conversations written in one exact format, with its own markers for where each message starts and ends. If your prompt is off by one marker, the model is reading something it was never trained on. We rebuilt the format of 740 models in Affordance and checked each one against the official version, token for token.

Models

740

Test conversations through each

963

Tokens

matched every time
Qwen · 120CodeQwen1.5-7B-ChatQwen-AgentWorld-35B-A3BQwen-Drive-1.0-4BQwen-Image-BenchQwen1.5-0.5Band 115 more
deepseek-ai · 40deepseek-coder-1.3b-instructdeepseek-coder-33b-instructdeepseek-coder-6.7b-instructdeepseek-coder-7b-instruct-v1.5DeepSeek-Coder-V2-Baseand 35 more
nvidia · 36Cosmos-Reason1-7Bgpt-oss-puzzle-88BLlama-3.1-Nemotron-70B-Instruct-HFLlama-3.1-Nemotron-8B-UltraLong-4M-InstructLlama-3.1-Nemotron-Nano-4B-v1.1and 31 more
LiquidAI · 29d1-2Bd1-3BLFM2-1.2BLFM2-2.6BLFM2-2.6B-Expand 24 more
google · 27codegemma-7b-itdiffusiongemma-26B-A4B-itfunctiongemma-270m-itgemma-1.1-2b-itgemma-1.1-7b-itand 22 more
microsoft · 27bitnet-b1.58-2B-4TDialoGPT-mediumDialoGPT-smallFara-7BFara1.5-27Band 22 more
zai-org · 23AutoGLM-Phone-9BAutoGLM-Phone-9B-Multilingualcogvlm2-llama3-chat-19BGLM-4-32B-0414GLM-4-9B-0414and 18 more
ibm-granite · 22granite-3.0-8b-instructgranite-3.1-8b-instructgranite-3.2-8b-instructgranite-3.3-8b-instructgranite-4.0-3b-visionand 17 more
openbmb · 20AgentCPM-ExploreAgentCPM-ReportBitCPM-CANN-8BMiniCPM-2B-sft-bf16MiniCPM-Llama3-V-2_5and 15 more
meta-llama · 17Llama-2-13b-chat-hfLlama-2-70b-chat-hfLlama-2-7b-chat-hfLlama-3.1-405B-InstructLlama-3.1-70B-Instructand 12 more
NousResearch · 15Arete-27BDeepHermes-3-Llama-3-8B-PreviewHermes-2-Pro-Llama-3-8BHermes-2-Theta-Llama-3-8BHermes-3-Llama-3.1-70Band 10 more
inclusionAI · 14Ling-1TLing-2.6-1TLing-2.6-flashLing-3.0-flashLing-3.0-flash-Finand 9 more
moonshotai · 14Kimi-Dev-72BKimi-K2-BaseKimi-K2-InstructKimi-K2-Instruct-0905Kimi-K2-Thinkingand 9 more
allenai · 11Llama-3.1-Tulu-3-8BMolmo-7B-D-0924Molmo2-8BOLMo-2-0325-32B-InstructOlmo-3-32B-Thinkand 6 more
mistralai · 10Magistral-Small-2509Ministral-3-8B-Instruct-2512Mistral-7B-Instruct-v0.1Mistral-7B-Instruct-v0.2Mistral-7B-Instruct-v0.3and 5 more
tencent · 10Hunyuan-1.8B-InstructHunyuan-A13B-InstructHunyuanOCRHy-Memory-30B-A3BHy3and 5 more

These are the 16 publishers with the most models, out of 164 publishers. The number beside each name is how many of its models are included. The table of all 740 models is at the end of this page.

What a prompt format is

You send a model a list of messages. The model reads one long run of tokens, which are numbers that each stand for a piece of a word. Something has to turn the first into the second, and every model family does it differently. Here is the same short conversation in the format of three models:

Qwen3.6
<|im_start|>system
# Tools

GLM-5.3
[gMASK]<sop><|system|>Reasoning Effort: Max<|system|>
# Tools

Llama 3.1
<|begin_of_text|><|start_header_id|>system<|end_header_id|>

The model's authors publish the format as a chat template, a small program in a template language called Jinja. They also publish a tokenizer, the program that splits text into tokens. Running one and then the other gives the official tokens.

What we built and how we checked it

We wrote 57 formats by hand, each as an Affordance app. Another 135 models use one of those 57 unchanged: a fine-tuned model is a copy of a model that someone trained further, and it usually keeps the original's template. For the other 548 we wrote a converter, a program that reads a model's official chat template and writes the same format as an Affordance app.

Then we took 963 test conversations: plain chats, system messages, tools the model can call, the results of those calls, and the model's own reasoning. For each model we ran every conversation through its app and through the model's official template, both with that model's own tokenizer, and compared the two lists of tokens.

Across all 740 modelsConversations
Our app and the official tools gave the same tokens539,564
The official template rejected the conversation, and so did our app173,056
Different0
Compared712,620

A template can reject a conversation it does not support. A model with no tool format rejects any conversation that uses tools, and many of the test conversations use tools, which is why about a quarter of the comparisons end this way. For those we checked that our app rejects the same conversations.

Why it is useful

These apps are what Affordance uses when you write an agent and name a model. You name the model the way its publisher does, for example moonshotai/Kimi-K2-Instruct, and the agent's prompt comes out in that model's format. They are also what we put inside SGLang and vLLM in place of the Jinja template.

If your model is not in the table, one command adds it. It fetches the model's template, converts it, and registers the model only if every test conversation gives the official tokens. The guide to adding a model has the command and what to do if it reports a difference.

The apps are in the repository under views/. If a token ever differs from your model's template, open an issue and we will fix it.

Every model

Each row is one model, under its publisher's name on Hugging Face. Every model went through 963 conversations.

ModelSame tokensRejected by bothDifferent
Qwen
CodeQwen1.5-7B-Chat6503130
Qwen-AgentWorld-35B-A3B7402230
Qwen-Drive-1.0-4B7402230
Qwen-Image-Bench7402230
Qwen1.5-0.5B6503130
Qwen1.5-0.5B-Chat6503130
Qwen1.5-110B-Chat6503130
Qwen1.5-14B-Chat6503130
Qwen1.5-32B6503130
Qwen1.5-32B-Chat6503130
Qwen1.5-72B-Chat6503130
Qwen1.5-7B-Chat6503130
Qwen1.5-MoE-A2.7B6503130
Qwen1.5-MoE-A2.7B-Chat6503130
Qwen2-0.5B6503130
Qwen2-0.5B-Instruct6503130
Qwen2-1.5B6503130
Qwen2-1.5B-Instruct6503130
Qwen2-72B6503130
Qwen2-72B-Instruct6503130
Qwen2-7B6503130
Qwen2-7B-Instruct6503130
Qwen2-VL-2B-Instruct864990
Qwen2-VL-72B-Instruct864990
Qwen2-VL-7B-Instruct864990
Qwen2.5-0.5B7232400
Qwen2.5-0.5B-Instruct7232400
Qwen2.5-1.5B7232400
Qwen2.5-1.5B-Instruct7232400
Qwen2.5-14B7232400
Qwen2.5-14B-Instruct7232400
Qwen2.5-14B-Instruct-1M7232400
Qwen2.5-32B7232400
Qwen2.5-32B-Instruct7232400
Qwen2.5-3B7232400
Qwen2.5-3B-Instruct7232400
Qwen2.5-72B7232400
Qwen2.5-72B-Instruct7232400
Qwen2.5-7B7232400
Qwen2.5-7B-Instruct7232400
Qwen2.5-7B-Instruct-1M7232400
Qwen2.5-Coder-0.5B-Instruct7232400
Qwen2.5-Coder-1.5B7232400
Qwen2.5-Coder-1.5B-Instruct7232400
Qwen2.5-Coder-14B7232400
Qwen2.5-Coder-14B-Instruct7232400
Qwen2.5-Coder-32B7232400
Qwen2.5-Coder-32B-Instruct7232400
Qwen2.5-Coder-3B-Instruct7232400
Qwen2.5-Coder-7B7232400
Qwen2.5-Coder-7B-Instruct7232400
Qwen2.5-Math-1.5B7232400
Qwen2.5-Math-7B7232400
Qwen2.5-Math-7B-Instruct7232400
Qwen2.5-Omni-3B864990
Qwen2.5-Omni-7B864990
Qwen2.5-VL-32B-Instruct864990
Qwen2.5-VL-3B-Instruct864990
Qwen2.5-VL-72B-Instruct864990
Qwen2.5-VL-7B-Instruct864990
Qwen3-0.6B932310
Qwen3-0.6B-Base6503130
Qwen3-1.7B932310
Qwen3-1.7B-Base6503130
Qwen3-14B932310
Qwen3-235B-A22B932310
Qwen3-235B-A22B-Instruct-2507932310
Qwen3-235B-A22B-Thinking-2507932310
Qwen3-30B-A3B932310
Qwen3-30B-A3B-Instruct-2507932310
Qwen3-30B-A3B-Thinking-2507932310
Qwen3-32B932310
Qwen3-4B932310
Qwen3-4B-Base6503130
Qwen3-4B-Instruct-2507932310
Qwen3-4B-Thinking-2507932310
Qwen3-8B932310
Qwen3-8B-Base6503130
Qwen3-Coder-30B-A3B-Instruct6103530
Qwen3-Coder-480B-A35B-Instruct6103530
Qwen3-Coder-Next6103530
Qwen3-Next-80B-A3B-Instruct932310
Qwen3-Next-80B-A3B-Thinking932310
Qwen3-Omni-30B-A3B-Captioner8571060
Qwen3-Omni-30B-A3B-Instruct864990
Qwen3-Omni-30B-A3B-Thinking8571060
Qwen3-VL-235B-A22B-Instruct864990
Qwen3-VL-235B-A22B-Thinking864990
Qwen3-VL-2B-Instruct864990
Qwen3-VL-2B-Thinking864990
Qwen3-VL-30B-A3B-Instruct864990
Qwen3-VL-30B-A3B-Thinking864990
Qwen3-VL-32B-Instruct864990
Qwen3-VL-32B-Thinking864990
Qwen3-VL-4B-Instruct864990
Qwen3-VL-4B-Thinking864990
Qwen3-VL-8B-Instruct864990
Qwen3-VL-8B-Thinking864990
Qwen3.5-0.8B7402230
Qwen3.5-0.8B-Base7402230
Qwen3.5-122B-A10B7402230
Qwen3.5-27B7402230
Qwen3.5-2B7402230
Qwen3.5-2B-Base7402230
Qwen3.5-35B-A3B7402230
Qwen3.5-35B-A3B-Base7402230
Qwen3.5-397B-A17B7402230
Qwen3.5-4B7402230
Qwen3.5-4B-Base7402230
Qwen3.5-9B7402230
Qwen3.5-9B-Base7402230
Qwen3.6-27B7402230
Qwen3.6-35B-A3B7402230
Qwen3.8-2.4T-A95B4145490
Qwen3.8-27B6383250
Qwen3.8-Flash-Next6383250
Qwen3Guard-Gen-0.6B6503130
Qwen3Guard-Gen-8B6503130
QwQ-32B6563070
QwQ-32B-Preview7232400
deepseek-ai
deepseek-coder-1.3b-instruct6682950
deepseek-coder-33b-instruct6682950
deepseek-coder-6.7b-instruct6682950
deepseek-coder-7b-instruct-v1.56682950
DeepSeek-Coder-V2-Base6503130
DeepSeek-Coder-V2-Instruct6503130
DeepSeek-Coder-V2-Lite-Base6503130
DeepSeek-Coder-V2-Lite-Instruct6503130
deepseek-llm-67b-chat6503130
deepseek-llm-7b-chat6503130
deepseek-math-7b-instruct6503130
deepseek-moe-16b-chat6503130
DeepSeek-R14714920
DeepSeek-R1-05284704930
DeepSeek-R1-0528-Qwen3-8B4704930
DeepSeek-R1-Distill-Llama-70B6732900
DeepSeek-R1-Distill-Llama-8B6732900
DeepSeek-R1-Distill-Qwen-1.5B6732900
DeepSeek-R1-Distill-Qwen-14B6732900
DeepSeek-R1-Distill-Qwen-32B6732900
DeepSeek-R1-Distill-Qwen-7B6732900
DeepSeek-R1-Zero4714920
DeepSeek-V26503130
DeepSeek-V2-Chat6503130
DeepSeek-V2-Chat-06286503130
DeepSeek-V2-Lite6503130
DeepSeek-V2-Lite-Chat6503130
DeepSeek-V2.56732900
DeepSeek-V36543090
DeepSeek-V3-03244714920
DeepSeek-V3.14714920
DeepSeek-V3.1-Base4714920
DeepSeek-V3.1-Terminus4714920
DeepSeek-V3.25524110
DeepSeek-V3.2-Exp4714920
DeepSeek-V4-Flash7352280
DeepSeek-V4-Pro7352280
DeepSeek-V4.1-Flash8271360
DeepSeek-V4.1-Flash-10177941690
deepseek-vl2-small6682950
nvidia
Cosmos-Reason1-7B864990
gpt-oss-puzzle-88B6153480
Llama-3.1-Nemotron-70B-Instruct-HF7372260
Llama-3.1-Nemotron-8B-UltraLong-4M-Instruct7372260
Llama-3.1-Nemotron-Nano-4B-v1.16623010
Llama-3.1-Nemotron-Nano-8B-v16972660
Llama-3.1-Nemotron-Nano-VL-8B-V1864990
Llama-3_1-Nemotron-Ultra-253B-v1864990
Llama-3_3-Nemotron-Super-49B-v1864990
Llama-3_3-Nemotron-Super-49B-v1_58401230
Llama3-ChatQA-1.5-8B6682950
LocateAnything-3B7232400
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF167941690
Nemotron-3.5-Parse96300
Nemotron-Cascade-2-30B-A3B7691940
Nemotron-Labs-Diffusion-14B7691940
Nemotron-Mini-4B-Instruct6503130
Nemotron-Orchestrator-8B932310
NV-Reason-CT7402230
NVIDIA-Nemotron-3-Nano-30B-A3B-BF167691940
NVIDIA-Nemotron-3-Nano-4B-BF167691940
NVIDIA-Nemotron-3-Super-120B-A12B-BF167691940
NVIDIA-Nemotron-3-Super-120B-A12B-NVFP47691940
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF167691940
NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP47691940
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF167691940
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP47691940
NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF167941690
NVIDIA-Nemotron-Labs-Dory1.0-4B-BF167691940
NVIDIA-Nemotron-Nano-12B-v26503130
NVIDIA-Nemotron-Nano-12B-v2-VL-BF16864990
NVIDIA-Nemotron-Nano-9B-v26503130
NVIDIA-Nemotron-Nano-9B-v2-Japanese6503130
NVIDIA-Nemotron-Parse-2.096300
NVIDIA-Nemotron-Parse-v1.196300
NVLM-D-72B7232400
LiquidAI
d1-2B8411220
d1-3B8411220
LFM2-1.2B8151480
LFM2-2.6B8151480
LFM2-2.6B-Exp8151480
LFM2-24B-A2B8411220
LFM2-350M8151480
LFM2-700M8151480
LFM2-8B-A1B8151480
LFM2-VL-1.6B864990
LFM2-VL-3B864990
LFM2-VL-450M864990
LFM2.5-1.2B-Base932310
LFM2.5-1.2B-Instruct8411220
LFM2.5-1.2B-JP932310
LFM2.5-1.2B-Thinking8411220
LFM2.5-2.6B8411220
LFM2.5-230M7432200
LFM2.5-24B-A2B8411220
LFM2.5-350M8411220
LFM2.5-8B-A1B7432200
LFM2.5-VL-1.6B864990
LFM2.5-VL-3B8411220
LFM2.5-VL-450M7432200
LFM3-1.7B8411220
LFM3-1.7B-Base8411220
LFM3-Audio-2B8411220
LFM3-Omni-2B8411220
LFM3-VL-2B8411220
google
codegemma-7b-it2187450
diffusiongemma-26B-A4B-it7941690
functiongemma-270m-it96300
gemma-1.1-2b-it2187450
gemma-1.1-7b-it2187450
gemma-2-27b-it2187450
gemma-2-2b-it2187450
gemma-2-2b-jpn-it2187450
gemma-2-9b-it2187450
gemma-3-12b-it5703930
gemma-3-1b-it5703930
gemma-3-270m-it5703930
gemma-3-27b-it5703930
gemma-3-4b-it5703930
gemma-3n-E2B-it5703930
gemma-3n-E4B-it5703930
gemma-4-12B-it7941690
gemma-4-26B-A4B-it7941690
gemma-4-31B-it7941690
gemma-4-E2B-it7941690
gemma-4-E4B-it7941690
recurrentgemma-2b-it2187450
shieldgemma-2-4b-it864990
shieldgemma-2b5014620
translategemma-12b-it09630
translategemma-27b-it09630
translategemma-4b-it09630
microsoft
bitnet-b1.58-2B-4T96300
DialoGPT-medium96300
DialoGPT-small96300
Fara-7B864990
Fara1.5-27B7402230
llava-med-v1.5-mistral-7b1647990
Mage-VL864990
Mage-VL-Cowork864990
MAI-DS-R14714920
Phi-3-medium-128k-instruct6682950
Phi-3-medium-4k-instruct6682950
Phi-3-mini-128k-instruct6503130
Phi-3-mini-4k-instruct6503130
Phi-3-small-128k-instruct6503130
Phi-3-small-8k-instruct6503130
Phi-3-vision-128k-instruct6503130
Phi-3.5-mini-instruct6503130
Phi-3.5-MoE-instruct6503130
Phi-3.5-vision-instruct6503130
phi-46503130
Phi-4-mini-flash-reasoning6503130
Phi-4-mini-instruct6503130
Phi-4-mini-reasoning6682950
Phi-4-reasoning6682950
Phi-4-reasoning-plus6682950
Phi-4-reasoning-vision-15B6682950
UserLM-8b96300
zai-org
AutoGLM-Phone-9B96300
AutoGLM-Phone-9B-Multilingual96300
cogvlm2-llama3-chat-19B96300
GLM-4-32B-041496300
GLM-4-9B-041496300
glm-4-9b-chat-hf96300
GLM-4.1V-9B-Thinking96300
GLM-4.57941690
GLM-4.5-Air7941690
GLM-4.5V7941690
GLM-4.67941690
GLM-4.6V7941690
GLM-4.6V-Flash7941690
GLM-4.77941690
GLM-4.7-Flash7941690
GLM-57941690
GLM-5.17941690
GLM-5.27941690
GLM-5.37941690
GLM-5.3-Flash7941690
GLM-Z1-32B-04146682950
GLM-Z1-9B-04146682950
GLM-Z1-Rumination-32B-04146682950
ibm-granite
granite-3.0-8b-instruct6503130
granite-3.1-8b-instruct6503130
granite-3.2-8b-instruct6503130
granite-3.3-8b-instruct6503130
granite-4.0-3b-vision864990
granite-4.0-h-1b96300
granite-4.0-h-350m96300
granite-4.0-h-micro96300
granite-4.0-h-small96300
granite-4.0-h-tiny96300
granite-4.0-micro96300
granite-4.0-tiny-preview6503130
granite-4.1-30b96300
granite-4.1-3b96300
granite-4.1-8b96300
granite-4.2-30b7691940
granite-4.2-3b7691940
granite-4.2-8b7691940
granite-docling-258M864990
granite-vision-3.1-2b-preview6503130
granite-vision-3.2-2b6503130
granite-vision-4.1-4b864990
openbmb
AgentCPM-Explore932310
AgentCPM-Report6503130
BitCPM-CANN-8B6503130
MiniCPM-2B-sft-bf167352280
MiniCPM-Llama3-V-2_596300
MiniCPM-o-2_67232400
MiniCPM-o-4_56503130
MiniCPM-OCR-1.07402230
MiniCPM-OCR-Jev7402230
MiniCPM-SALA7691940
MiniCPM-V-46503130
MiniCPM-V-4.67402230
MiniCPM-V-4.7-1B7402230
MiniCPM-V-4.7-35B-A3B7402230
MiniCPM-V-4_56503130
MiniCPM3-4B6103530
MiniCPM4-8B6503130
MiniCPM4.1-8B6503130
MiniCPM5-1B7691940
MiniCPM5-2B7691940
meta-llama
Llama-2-13b-chat-hf4375260
Llama-2-70b-chat-hf4375260
Llama-2-7b-chat-hf4375260
Llama-3.1-405B-Instruct7372260
Llama-3.1-70B-Instruct7372260
Llama-3.1-8B-Instruct7372260
Llama-3.2-11B-Vision-Instruct7122510
Llama-3.2-1B-Instruct7372260
Llama-3.2-3B-Instruct7372260
Llama-3.3-70B-Instruct7372260
Llama-4-Maverick-17B-128E-Instruct864990
Llama-4-Scout-17B-16E-Instruct864990
Llama-Guard-3-1B2187450
Llama-Guard-3-8B1677960
Llama-Guard-4-12B2187450
Meta-Llama-3-70B-Instruct96300
Meta-Llama-3-8B-Instruct96300
NousResearch
Arete-27B6383250
DeepHermes-3-Llama-3-8B-Preview7082550
Hermes-2-Pro-Llama-3-8B5793840
Hermes-2-Theta-Llama-3-8B5793840
Hermes-3-Llama-3.1-70B5793840
Hermes-3-Llama-3.1-8B5793840
Hermes-3-Llama-3.2-3B6503130
Hermes-4-14B6533100
Hermes-4-405B6533100
Hermes-4-70B6533100
Hermes-4.3-36B6533100
Meta-Llama-3-8B-Instruct96300
Nous-Hermes-2-Mixtral-8x7B-DPO6503130
Nous-Hermes-2-SOLAR-10.7B6503130
Nous-Hermes-2-Yi-34B6503130
inclusionAI
Ling-1T7132500
Ling-2.6-1T7132500
Ling-2.6-flash7132500
Ling-3.0-flash6103530
Ling-3.0-flash-Fin6103530
Ling-3.0-flash-VL7941690
Ling-3.0-tiny6103530
Ling-3.1-flash6103530
Ling-3.1-flash-10196103530
Ling-flash-2.07132500
Ling-mini-2.07132500
LLaDA2.1-flash7132500
LLaDA2.1-mini7132500
Ring-2.6-1T6283350
moonshotai
Kimi-Dev-72B7232400
Kimi-K2-Base864990
Kimi-K2-Instruct96300
Kimi-K2-Instruct-090596300
Kimi-K2-Thinking96300
Kimi-K2.596300
Kimi-K2.696300
Kimi-K2.7-Code96300
Kimi-Linear-48B-A3B-Instruct96300
Kimi-VL-A3B-Instruct864990
Kimi-VL-A3B-Thinking864990
Kimi-VL-A3B-Thinking-2506864990
Moonlight-16B-A3B96300
Moonlight-16B-A3B-Instruct96300
allenai
Llama-3.1-Tulu-3-8B6503130
Molmo-7B-D-09241647990
Molmo2-8B2147490
OLMo-2-0325-32B-Instruct6503130
Olmo-3-32B-Think7132500
Olmo-3-7B-Instruct6103530
Olmo-3-7B-Think7132500
Olmo-3.1-32B-Think7132500
olmOCR-2-7B-1025864990
olmOCR-7B-0225-preview864990
OLMoE-1B-7B-0924-Instruct6503130
mistralai
Magistral-Small-25098311320
Ministral-3-8B-Instruct-2512927360
Mistral-7B-Instruct-v0.14285350
Mistral-7B-Instruct-v0.24285350
Mistral-7B-Instruct-v0.36313320
Mistral-Large-4-1T-A52B7622010
Mistral-Medium-3.5-128B6982650
Mistral-Nemo-Instruct-24076313320
Mistral-Small-3.2-24B-Instruct-25064205430
Mistral-Small-4-119B-26036982650
tencent
Hunyuan-1.8B-Instruct7132500
Hunyuan-A13B-Instruct6343290
HunyuanOCR96300
Hy-Memory-30B-A3B7941690
Hy36183450
Hy3-preview7941690
Hy4-preview5723910
WeVisOne-VL-4B864990
Youtu-LLM-2B6503130
Youtu-Parsing-Omni864990
huihui-ai
Huihui-gemma-4-12B-it-abliterated96300
Huihui-Qwen3-VL-4B-Instruct-abliterated864990
Huihui-Qwen3-VL-8B-Instruct-abliterated864990
Huihui-Qwen3.5-27B-abliterated7402230
Huihui-Qwen3.5-9B-abliterated7402230
Huihui-Qwen3.6-35B-A3B-abliterated7402230
Huihui-Qwen3.8-27B-abliterated6383250
Llama-3.2-3B-Instruct-abliterated7372260
Llama-3.3-70B-Instruct-abliterated7372260
MiniMaxAI
MiniMax-M1-40k69570
MiniMax-M27941690
MiniMax-M2.17941690
MiniMax-M2.57941690
MiniMax-M2.77941690
MiniMax-M37941690
MiniMax-M3.1-Flash7941690
MiniMax-Text-0189550
MiniMax-VL-01864990
XiaomiMiMo
MiMo-7B-Base7232400
MiMo-7B-RL7232400
MiMo-V2-Flash6103530
MiMo-V2.57432200
MiMo-V2.5-Pro7432200
MiMo-V2.6-Distill-Qwen-9B96300
MiMo-V2.6-Flash-RL96300
MiMo-V2.6-Pro-RL96300
MiMo-VL-7B-RL-25087232400
LGAI-EXAONE
EXAONE-3.5-2.4B-Instruct6503130
EXAONE-3.5-32B-Instruct6503130
EXAONE-3.5-7.8B-Instruct6503130
EXAONE-4.0-1.2B96300
EXAONE-4.0-32B96300
EXAONE-4.5-33B96300
K-EXAONE-2.0-750B-A37B7941690
K-EXAONE-236B-A23B96300
ATH-MaaS
Ovis1.6-Gemma2-9B2187450
Ovis2-1B7232400
Ovis2.5-2B5973660
Ovis2.5-9B5973660
Ovis2.6-30B-A3B5973660
OvisOCR27402230
OvisOCR37402230
ByteDance-Seed
BAGEL-7B-MoT7232400
Seed-Coder-8B-Instruct96300
Seed-Coder-8B-Reasoning2187450
Seed-OSS-36B-Instruct5953680
UI-TARS-1.5-7B7232400
UI-TARS-72B-DPO96300
UI-TARS-7B-DPO96300
ornith-ai
Ornith-1.0-35B7402230
Ornith-1.0-397B7402230
Ornith-1.0-9B7402230
Ornith-1.5-122B7402230
Ornith-1.5-35B-A3B7402230
Ornith-1.5-397B7402230
Ornith-1.5-9B7402230
baidu
ERNIE-4.5-0.3B-PT6503130
ERNIE-4.5-21B-A3B-PT6503130
ERNIE-4.5-21B-A3B-Thinking932310
ERNIE-4.5-VL-28B-A3B-PT96300
Qianfan-OCR864990
Unlimited-OCR96300
internlm
Intern-Decision-4B7402230
Intern-S1864990
Intern-S1-mini864990
Intern-S1-Pro864990
internlm2_5-7b-chat6503130
internlm3-8b-instruct6503130
arcee-ai
AFM-4.5B6503130
Llama-3.1-SuperNova-Lite96300
Trinity-Large-Preview932310
Trinity-Large-Thinking96300
Trinity-Mini932310
dphn
dolphin-2.6-mixtral-8x7b6503130
dolphin-2.9-llama3-8b6503130
dolphin-2.9.3-mistral-nemo-12b6503130
Dolphin-Mistral-24B-Venice-Edition951120
Dolphin3.0-Mistral-24B7232400
IFM
K2-Horizon-0.9B1338300
K2-Horizon-3.7B1338300
K2-Horizon-375B-A23B1338300
K2-Horizon-7B1338300
K2-Horizon-MoVA-36B-A4B1338300
stepfun-ai
Step-3.5-Flash7941690
Step-3.7-Flash7941690
Step-Audio-2-mini7232400
step37941690
Step3-VL-10B96300
01-ai
Yi-1.5-34B-Chat6682950
Yi-1.5-9B-Chat6682950
Yi-34B-Chat6503130
Yi-Coder-9B-Chat6503130
mlabonne
AlphaMonarch-7B6503130
gemma-3-27b-it-abliterated5703930
Meta-Llama-3.1-8B-Instruct-abliterated96300
NeuralDaredevil-8B-abliterated96300
Nanbeige
Nanbeige4-3B-Thinking-2511932310
Nanbeige4.1-3B932310
Nanbeige4.2-3B7402230
Nanbeige4.3-33B7402230
openai
gpt-oss-120b6153480
gpt-oss-20b6153480
gpt-oss-safeguard-120b6153480
gpt-oss-safeguard-20b6153480
poolside
Laguna-M.17691940
Laguna-S-2.17691940
Laguna-XS-2.17691940
Laguna-XS.27691940
sarvamai
sarvam-14285350
sarvam-105b7941690
sarvam-30b7941690
sarvam-m4285350
skt
A.X-4.06862770
A.X-4.0-Light6862770
A.X-K16862770
A.X-K296300
tiiuae
falcon-11B6503130
falcon-7b-instruct4375260
Falcon-H1R-7B6682950
Falcon3-10B-Instruct7232400
upstage
SOLAR-10.7B-Instruct-v1.06503130
Solar-Open-100B7352280
Solar-Open2-250B6283350
solar-pro-preview-instruct6503130
ByteDance
Ouro-1.4B6503130
Ouro-2.6B6503130
Ouro-2.6B-Thinking6503130
Cloudflare
clef6383250
clef-flash7402230
clef-omni864990
codellama
CodeLlama-13b-Instruct-hf5314320
CodeLlama-34b-Instruct-hf5314320
CodeLlama-7b-Instruct-hf5314320
CohereLabs
command-a-plus-05-2026-bf16924390
North-Micro-Vision-Instruct864990
North-Mini-Code-1.0924390
datalab-to
chandra-ocr-27402230
lift5584050
surya-ocr-25584050
DavidAU
Mistral-Nemo-2407-12B-Thinking-Claude-Gemini-GPT5.2-Uncensored-HERETIC4325310
Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU6383250
Qwen3.8-27B-Ultimate-Heretic-Uncensored6543090
dots-studio
dots.mocr6503130
dots.ocr6503130
dots3-note-prev7941690
empero-ai
Qwen3.8-35B-A3B-Distill7402230
Qwen3.8-9B-Distill7402230
Qwythos-9B-Claude-Mythos-5-1M7402230
HuggingFaceTB
SmolLM2-135M-Instruct6503130
SmolLM3-3B932310
SmolVLM2-2.2B-Instruct864990
InternScience
Agents-A17402230
Agents-A1-4B7402230
Sage-A1.57402230
lightonai
LightOnOCR-2-1B864990
LightOnOCR-3-0.8B7402230
LightOnOCR-3-4B7402230
llava-hf
llava-1.5-7b-hf96300
llava-v1.6-34b-hf6503130
llava-v1.6-mistral-7b-hf1647990
meituan-longcat
LongCat-2.07292340
LongCat-Flash-Chat8181450
LongCat-Flash-Lite7062570
naver-hyperclovax
HyperCLOVAX-SEED-Omni-8B7552080
HyperCLOVAX-SEED-Think-14B96300
HyperCLOVAX-SEED-Think-32B7941690
openchat
openchat-3.5-01066503130
openchat-3.5-12106503130
openchat-3.6-8b-202405225973660
opendatalab
MinerU2.5-2509-1.2B864990
MinerU2.5-Pro-2604-1.2B864990
MinerU2.5-Pro-2605-1.2B864990
OpenGVLab
InternVL3-78B8401230
InternVL3-8B8401230
InternVL3_5-8B864990
PaddlePaddle
PaddleOCR-VL864990
PaddleOCR-VL-1.5864990
PaddleOCR-VL-1.6864990
thinkingmachines
Inkling7552080
Inkling-2-Small7941690
Inkling-Small7552080
ukisai
Swift-1.5-Qwen3.8-27b6383250
Swift-Qwen3.8-27b6383250
Swift1.5-Qwen3.8-Flash-Next6383250
XingChen-AGI
TeleOCR864990
Xing4.0-29B-A4B7941690
Xing4.0-29B-Intent7941690
AEON-7
Qwen3.6-27B-AEON-Ultimate-Uncensored-BF167402230
Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-BF166383250
baichuan-inc
Baichuan-M2-32B6503130
Baichuan-M3-235B932310
GSAI-ML
LLaDA-8B-Base96300
LLaDA-8B-Instruct96300
HuggingFaceH4
zephyr-7b-alpha6503130
zephyr-7b-beta6503130
JetBrains
Mellum2-12B-A2.5B-Thinking96300
Mellum2.1-12B-A2.5B-Thinking96300
liuhaotian
llava-v1.6-34b6503130
llava-v1.6-mistral-7b1647990
manycore-research
SpatialLM-Llama-1B7372260
SpatialLM-Qwen-0.5B7232400
nanonets
Nanonets-OCR-s864990
Nanonets-OCR2-3B864990
Nexusflow
Athene-70B96300
Athene-V2-Chat7232400
Sao10K
L3-8B-Lunaris-v196300
L3-8B-Stheno-v3.296300
sensenova
SenseNova-Loop-9B-Preview7402230
SenseNova-U1.5-8B-MoT6503130
swiss-ai
Apertus-70B-Instruct-25093476160
Apertus-8B-Instruct-25093476160
XHToken
Spark-X2.5-4B7192440
Spark-X2.6-4B5933700
5CD-AI
Vintern-1B-v3_57232400
abacusai
Smaug-Llama-3-70B-Instruct96300
agentica-org
DeepScaleR-1.5B-Preview6732900
ai21labs
AI21-Jamba-Reasoning-3B6503130
Aleph-Alpha
Kolibri-17352280
Alibaba-NLP
Tongyi-DeepResearch-30B-A3B932310
alpindale
WizardLM-2-8x22B6503130
apple
FastVLM-0.5B6503130
AtlaAI
Selene-1-Mini-Llama-3.1-8B7372260
autotrust
Qwen3.8-27B-NVFP4-DGX-SPARK-V26383250
BAAI
AREX-26383250
beowolx
CodeNinja-1.0-OpenChat-7B6503130
BioMistral
BioMistral-7B1647990
bosonai
Higgs-Llama-3-70B96300
chatdb
natural-sql-7b6682950
chuanli11
Llama-3.2-3B-Instruct-uncensored7372260
coder3101
gemma-4-31B-it-heretic96300
dealignai
Gemma-4-31B-JANG_4M-CRACK96300
Deci
DeciLM-7B-instruct6503130
DeepHat
DeepHat-V1-7B96300
Diffuspace
Clover-35B-A3B-Preview6712920
Dream-org
Dream-v0-Instruct-7B7232400
elyza
Llama-3-ELYZA-JP-8B96300
EssentialAI
rnj-1-instruct96300
failspy
llama-3-70B-Instruct-abliterated96300
fancyfeast
llama-joycaption-beta-one-hf-llava6682950
fastino
GLiDE-27B6383250
gradientai
Llama-3-8B-Instruct-Gradient-1048k96300
GritLM
GritLM-7B6682950
GrootTeam
Groot-19B-A3B-2610237402230
h2oai
h2o-lightning-4b7402230
Hcompany
Holo4-27B6383250
HemmingwayAI
Hemmingway-16383250
HuggingFaceBio
Carbon-X932310
HuggingFaceM4
Idefics3-8B-Llama396300
huggyllama
llama-7b4375260
humain-ai
ALLaM-7B-Instruct-preview4375260
IlyaGusev
saiga_llama3_8b96300
infly
Infinity-Parser2-Pro7402230
IQuestLab
IQuest-Q17941690
Jackrong
Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled932310
JonathanColetti
Qwen3.8-27B-Uncensored6383250
jondurbin
bagel-dpo-34b-v0.26503130
jzhang533
PaddleOCR-VL-For-Manga864990
K-intelligence
Midm-2.0-Base-Instruct96300
katanemo
Arch-Router-1.5B7232400
KRAFTON
Raon-1-mini-preview7192440
Kwaipilot
KAT-Coder-V2.5-Dev7941690
lightblue
suzume-llama-3-8B-multilingual96300
llamaindex
vdr-2b-multi-v1864990
llmfan46
gemma-4-31B-it-uncensored-heretic7941690
lukey03
Qwen3.5-9B-abliterated7402230
MemTensor
MemT-1.0-Nano7402230
meta-models
Muse-Glimmer-30B7292340
mistral-experimental
pixtral-12b5474160
Motif-Technologies
Motif-396300
NaiveAI
Naive-N0.5-Flash7432200
nex-agi
Nex-N2.5-mini7941690
OBLITERATUS
DeepSeek-V4.1-Flash-1017-obliterated96300
Orenguteng
Llama-3.1-8B-Lexi-Uncensored-V27372260
p-e-w
gpt-oss-20b-heretic6153480
paradigma-inc
limite-1b-violetto1388250
PerceptronAI
Isaac-0.1932310
PleIAs
Baguettotron96300
project-numina
NuminaMath-7B-TIR6682950
prometheus-eval
prometheus-7b-v2.04285350
qnguyen3
nanoLLaVA6503130
RecursiveSI
flint7941690
reflection
Beam-1.0-501B-A23B6103530
refuelai
Llama-3-Refueled96300
rhymes-ai
Aria96300
RikiyaT
Qwen3.5-0.8B-weight-space-programming7402230
ServiceNow-AI
Apriel-Nemotron-15b-Thinker7132500
shenzhi-wang
Llama3-8B-Chinese-Chat96300
Skywork
Skywork-R1V-38B6732900
Snowflake
snowflake-arctic-instruct6503130
stabilityai
stablelm-zephyr-3b6503130
taobao-mnn
MNN-Pocket-0.3B932310
teknium
OpenHermes-2.5-Mistral-7B6503130
TinyLlama
TinyLlama-1.1B-Chat-v1.06503130
TokenRhythm
NeoHorse-1-4B7402230
tomg-group-umd
huginn-012596300
Tongyi-MAI
MAI-UI-8B864990
utter-project
EuroLLM-1.7B-Instruct96300
Vikhrmodels
Vikhr-Nemo-12B-Instruct-R-21-09-2496300
WeiboAI
VibeThinker-3B7232400
xiaomi-research
Xiaomi-OCR-07402230
xtools-at
DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic96300
YannQi
R-4B96300
yanolja
YanoljaNEXT-EEVE-Instruct-10.8B6503130
ZJU-AI4H
Hulu-Med-7B864990
Zyphra
Zamba2-7B-Instruct6503130