Open ASR Leaderboard
π
1.39k
Compare speech-to-text models by WER and speed
Mike Brummett
GoDjMike
AI & ML interests
Edge detection, road anomaly identification, story-generation libraries
Recent Activity
liked a Space about 2 hours ago
hf-audio/open_asr_leaderboard liked a model about 2 hours ago
Qwen/Qwen3-ASR-0.6B liked a model about 1 month ago
skilledu/Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW-i1-GGUFOrganizations
None yet
liked a
Space about 2 hours ago
liked a
model about 2 hours ago
liked 8
models about 1 month ago
skilledu/Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW-i1-GGUF
9B β’ Updated β’ 2.98k β’ 4
filvyb/Qwen3.6-9B-Heretic-Uncensored-Thinking-Sweet-Madness-GGUF
Image-Text-to-Text β’ 9B β’ Updated β’ 5.28k β’ 7
DavidAU/Qwen3.6-9B-Heretic-Uncensored-Thinking-Sweet-Madness
Image-Text-to-Text β’ 9B β’ Updated β’ 1.4k β’ 19
DavidAU/OpenAi-GPT-oss-20b-HERETIC-uncensored-NEO-Imatrix-gguf
Text Generation β’ 21B β’ Updated β’ 25.9k β’ 157
mradermacher/Artemis-31B-v1h-GGUF-i1-GGUF
31B β’ Updated β’ 443 β’ 3
LiquidAI/LFM2.5-Audio-1.5B-ONNX
Text-to-Audio β’ Updated β’ 201 β’ 19
LiquidAI/LFM2.5-Audio-1.5B
Audio-to-Audio β’ 1B β’ Updated β’ 1.5k β’ 430
nvidia/personaplex-7b-v1
Audio-to-Audio β’ 8B β’ Updated β’ 371k β’ 2.58k
liked 5
models 10 months ago
CohereLabs/command-a-translate-08-2025
Text Generation β’ 111B β’ Updated β’ 31 β’ β’ 79
Qwen/Qwen-Image
Text-to-Image β’ Updated β’ 196k β’ β’ 2.52k
unsloth/gemma-3-270m-it-GGUF
Text Generation β’ 0.3B β’ Updated β’ 122k β’ 165
google/gemma-3-270m-it
Text Generation β’ 0.3B β’ Updated β’ 392k β’ 600
google/gemma-3-270m
Text Generation β’ 0.3B β’ Updated β’ 6.97M β’ 1.04k
liked a
Space 11 months ago
reacted to georgewritescode's post with π 11 months ago
Post
3465
Announcing Artificial Analysis Long Context Reasoning (AA-LCR), a new benchmark to evaluate long context performance through testing reasoning capabilities across multiple long documents (~100k tokens)
The focus of AA-LCR is to replicate real knowledge work and reasoning tasks, testing capability critical to modern AI applications spanning document analysis, codebase understanding, and complex multi-step workflows.
AA-LCR is 100 hard text-based questions that require reasoning across multiple real-world documents that represent ~100k input tokens. Questions are designed so answers cannot be directly found but must be reasoned from multiple information sources, with human testing verifying that each question requires genuine inference rather than retrieval.
Key takeaways:
β€ Todayβs leading models achieve ~70% accuracy: the top three places go to OpenAI o3 (69%), xAI Grok 4 (68%) and Qwen3 235B 2507 Thinking (67%)
β€π We also already have gpt-oss results! 120B performs close to o4-mini (high), in-line with OpenAI claims regarding model performance. We will be following up shortly with a Intelligence Index for the models.
β€ 100 hard text-based questions spanning 7 categories of documents (Company Reports, Industry Reports, Government Consultations, Academia, Legal, Marketing Materials and Survey Reports)
β€ ~100k tokens of input per question, requiring models to support a minimum 128K context window to score on this benchmark
β€ ~3M total unique input tokens spanning ~230 documents to run the benchmark (output tokens typically vary by model)
Weβre adding AA-LCR to the Artificial Analysis Intelligence Index, and taking the version number to v2.2. Artificial Analysis Intelligence Index v2.2 now includes: MMLU-Pro, GPQA Diamond, AIME 2025, IFBench, LiveCodeBench, SciCode and AA-LCR.
Link to dataset: ArtificialAnalysis/AA-LCR
The focus of AA-LCR is to replicate real knowledge work and reasoning tasks, testing capability critical to modern AI applications spanning document analysis, codebase understanding, and complex multi-step workflows.
AA-LCR is 100 hard text-based questions that require reasoning across multiple real-world documents that represent ~100k input tokens. Questions are designed so answers cannot be directly found but must be reasoned from multiple information sources, with human testing verifying that each question requires genuine inference rather than retrieval.
Key takeaways:
β€ Todayβs leading models achieve ~70% accuracy: the top three places go to OpenAI o3 (69%), xAI Grok 4 (68%) and Qwen3 235B 2507 Thinking (67%)
β€π We also already have gpt-oss results! 120B performs close to o4-mini (high), in-line with OpenAI claims regarding model performance. We will be following up shortly with a Intelligence Index for the models.
β€ 100 hard text-based questions spanning 7 categories of documents (Company Reports, Industry Reports, Government Consultations, Academia, Legal, Marketing Materials and Survey Reports)
β€ ~100k tokens of input per question, requiring models to support a minimum 128K context window to score on this benchmark
β€ ~3M total unique input tokens spanning ~230 documents to run the benchmark (output tokens typically vary by model)
Weβre adding AA-LCR to the Artificial Analysis Intelligence Index, and taking the version number to v2.2. Artificial Analysis Intelligence Index v2.2 now includes: MMLU-Pro, GPQA Diamond, AIME 2025, IFBench, LiveCodeBench, SciCode and AA-LCR.
Link to dataset: ArtificialAnalysis/AA-LCR
reacted to prithivMLmods's post with π 11 months ago
Post
4297
Qwen Image β The Latest Image Generation Modelπ₯
Below are some samples generated using the Qwen Image Diffusion Model. Qwen-Image, a 20B MMDiT model for next-generation text-to-image generation, preserves typographic details, layout coherence, and contextual harmony with stunning accuracy. It is especially strong at creating stunning graphic posters with native text. The model is now open-source. [ ππ ππ-πΈππππ : Qwen/Qwen-Image ]
β€· Try the Qwen Image demo here: https://huggingface.co/spaces/prithivMLmods/Qwen-Image-Diffusion
β€· Qwen-Image Technical Report : Qwen-Image Technical Report (2508.02324)
β€· Qwen Image [GitHub] : https://github.com/QwenLM/Qwen-Image
Even more impressively, it demonstrates a strong ability to understand images. The model supports a wide range of vision-related tasks such as object detection, semantic segmentation, depth and edge (Canny) estimation, novel view synthesis, and image super-resolution. While each task is technically distinct, they can all be viewed as advanced forms of intelligent image editing driven by deep visual understanding. Collectively, these capabilities position Qwen-Image as more than just a tool for generating appealing visuals, it serves as a versatile foundation model for intelligent visual creation and transformation, seamlessly blending language, layout, and imagery.
Qwen-Image uses a dual-stream MMDiT architecture with a frozen Qwen2.5-VL, VAE encoder, RMSNorm for QK-Norm, LayerNorm elsewhere, and a custom MSRoPE scheme for joint image-text positional encoding.
.
.
.
To know more about it, visit the model card of the respective model. !!
Below are some samples generated using the Qwen Image Diffusion Model. Qwen-Image, a 20B MMDiT model for next-generation text-to-image generation, preserves typographic details, layout coherence, and contextual harmony with stunning accuracy. It is especially strong at creating stunning graphic posters with native text. The model is now open-source. [ ππ ππ-πΈππππ : Qwen/Qwen-Image ]
β€· Try the Qwen Image demo here: https://huggingface.co/spaces/prithivMLmods/Qwen-Image-Diffusion
β€· Qwen-Image Technical Report : Qwen-Image Technical Report (2508.02324)
β€· Qwen Image [GitHub] : https://github.com/QwenLM/Qwen-Image
Even more impressively, it demonstrates a strong ability to understand images. The model supports a wide range of vision-related tasks such as object detection, semantic segmentation, depth and edge (Canny) estimation, novel view synthesis, and image super-resolution. While each task is technically distinct, they can all be viewed as advanced forms of intelligent image editing driven by deep visual understanding. Collectively, these capabilities position Qwen-Image as more than just a tool for generating appealing visuals, it serves as a versatile foundation model for intelligent visual creation and transformation, seamlessly blending language, layout, and imagery.
Qwen-Image uses a dual-stream MMDiT architecture with a frozen Qwen2.5-VL, VAE encoder, RMSNorm for QK-Norm, LayerNorm elsewhere, and a custom MSRoPE scheme for joint image-text positional encoding.
.
.
.
To know more about it, visit the model card of the respective model. !!
liked 2
models 11 months ago
