LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

New and trending LLMs

DeepSeek V4.1 Flash 1017
1,564
Muse Spark Open
535
LFM2.5-24B-A2B
485

Rising on Hugging Face · read October 20

  1. New DeepSeek V4.1 Flash 1017 October 17
  2. New E Glimpse October 17
  3. New Mercury 3 October 16
  4. New R Beam 1.0 501B A23B October 16

On October 20, 2026, DeepSeek V4.1 Flash 1017 had the highest Hugging Face trending score of any open-weight model in our catalog, 1,564. 43 models were released in the last 30 days, and 17 of them already have a coding score.

Released in the last 30 days: 43 models

Every release on the day OpenRouter lists it, newest first. 17 already have a coding index and a place on the ranking; the rest are named under each day. Release dates: OpenRouter

Best debut of the last 30 days

Claude Opus 5.5

Anthropic, released September 22

92.2 Vibe Coding Index, #1 of 136 on the coding ranking
  1. Claude Sonnet 5.5 88.9 #3 of 136, released September 28
  2. GPT-6.1 Sol 88.7 #4 of 136, released September 29

  1. 2 releases

    Not benchmarked yet E Glimpse

  2. 2 releases

    Not benchmarked yet Mercury 3

  3. 5 releases

    Not benchmarked yet M Moondream 3 Preview M Moondream 3.1 U Pareto 26.10

  4. 5 releases

    Not benchmarked yet LFM2.5-24B-A2B#3 on Hugging Face trending U Pareto 26.9 S Extract S Summarize

  5. 1 release

  6. 1 release

    • Step 5 Preview #26 of 136 in coding 71.4 Trending #8 on Hugging Face

  7. 1 release

  8. 2 releases

    Not benchmarked yet Nano Banana 2.1

  9. 1 release

    Not benchmarked yet Ling 3.1 Flash

  10. 2 releases

    • GPT-6.1 Sol #4 of 136 in coding 88.7 Top 10 in coding

    Not benchmarked yet GPT-6.1 Sol Pro

  11. 1 release

  12. 2 releases

    Not benchmarked yet Perceptron Mk1.5 T Jev Router

  13. 1 release

    Not benchmarked yet Ember-1

  14. 5 releases

    Not benchmarked yet A Aion 3.5 A Aion 3.5 Mini Qwen3.8 Max Prime Solar Mini 4 GLM 5.3 Prime

  15. 6 releases

    Not benchmarked yet Command A+ GPT-6 Luna Pro GPT-6 Sol Pro

  16. 6 releases

    Not benchmarked yet Switchyard Qwen3.8 Omni Flash MiMo-V2.6-Pro-UltraSpeed

All 43 as a table, with evidence and price
Models released in the last 30 days
ReleasedModelCoding indexEvidenceOutput $/Mtok
October 17 DeepSeek V4.1 Flash 1017DeepSeek 83.3 $0.60
October 17 E GlimpseElorian Not benchmarked yet $3.00
October 16 Mercury 3Inception Not benchmarked yet $0.35
October 16 R Beam 1.0 501B A23BReflection 33.8 $1.00
October 15 Muse Spark OpenMeta 62.1 $4.25
October 15 M Moondream 3 PreviewMoondream Not benchmarked yet $2.50
October 15 M Moondream 3.1Moondream Not benchmarked yet $2.50
October 15 Nemotron 3.5 Super VLNVIDIA 65.5 $0.50
October 15 U Pareto 26.10Unbiased Not benchmarked yet $3.20
October 13 Gemini 4 ArgonGoogle 80.5 $10.00
October 13 LFM2.5-24B-A2BLiquid Not benchmarked yet $0.70
October 13 U Pareto 26.9Unbiased Not benchmarked yet $7.50
October 13 S ExtractScaledown Not benchmarked yet Free
October 13 S SummarizeScaledown Not benchmarked yet Free
October 12 Muse WonderMeta 64.5 $20.00
October 8 Step 5 PreviewStepFun 71.4 $2.70
October 7 Claude Haiku 5.5Anthropic 75.9 $0.50
October 6 Nano Banana 2.1Google Not benchmarked yet $7.50
October 6 Mistral Large 4Mistral AI 55.3 $2.09
October 2 Ling 3.1 FlashinclusionAI Not benchmarked yet $0.225
September 29 GPT-6.1 SolOpenAI 88.7 $10.00
September 29 GPT-6.1 Sol ProOpenAI Not benchmarked yet $10.00
September 28 Claude Sonnet 5.5Anthropic 88.9 $10.00
September 25 Perceptron Mk1.5Perceptron Not benchmarked yet $1.50
September 25 T Jev RouterTypesafe Not benchmarked yet Not published
September 24 Ember-1Fireworks Not benchmarked yet $15.00
September 23 A Aion 3.5Aion Labs Not benchmarked yet $6.00
September 23 A Aion 3.5 MiniAion Labs Not benchmarked yet $1.40
September 23 Qwen3.8 Max PrimeQwen Not benchmarked yet $12.00
September 23 Solar Mini 4Upstage Not benchmarked yet $0.20
September 23 GLM 5.3 PrimeZ.ai Not benchmarked yet $8.80
September 22 Claude Opus 5.5Anthropic 92.2 $20.00
September 22 Command A+Cohere Not benchmarked yet $1.50
September 22 GPT-6 LunaOpenAI 71.4 $0.50
September 22 GPT-6 Luna ProOpenAI Not benchmarked yet $0.50
September 22 GPT-6 SolOpenAI 82.8 $10.00
September 22 GPT-6 Sol ProOpenAI Not benchmarked yet $10.00
September 21 SwitchyardNVIDIA Not benchmarked yet Not published
September 21 Qwen3.8 Omni FlashQwen Not benchmarked yet $0.47
September 21 Grok 4.7xAI 75.0 $6.00
September 21 MiMo-V2.6-FlashXiaomi 61.9 $0.28
September 21 MiMo-V2.6-ProXiaomi 72.6 $0.87
September 21 MiMo-V2.6-Pro-UltraSpeedXiaomi Not benchmarked yet $8.70

Rising on Hugging Face

Hugging Face's own trending score, which weighs recent likes. Downloads are Hugging Face's count for the last 30 days.

  1. 1 DeepSeek V4.1 Flash 1017 deepseek-ai/DeepSeek-V4.1-Flash-1017 1,724 likes 1,564 trending score 83.3coding
  2. 2 Meta Muse Spark Open meta-models/Muse-Spark-Open-816B-A42B 584 likes 535 trending score 62.1coding
  3. 3 LiquidAI LFM2.5-24B-A2B LiquidAI/LFM2.5-24B-A2B 501 likes 485 trending score Not benchmarked
  4. 4 Qwen3.8 27B Qwen/Qwen3.8-27B 18K likes 287 trending score 73.6coding
  5. 5 Qwen3.8 Flash Qwen/Qwen3.8-Flash-Next 6,336 likes 171 trending score 87.4coding
  6. 6 DeepSeek V4.1 Flash deepseek-ai/DeepSeek-V4.1-Flash 4,657 likes 169 trending score 77.2coding
  7. 7 R Reflection Beam 1.0 501B A23B reflection/Beam-1.0-501B-A23B 170 likes 168 trending score 33.8coding
  8. 8 StepFun Step 5 Preview stepfun-ai/Step-5-Preview 117 likes 116 trending score 71.4coding
  9. 9 P PrismML Ternary Bonsai 2 27B prism-ml/Ternary-Bonsai-2-27B-gguf 2,758 likes 77 trending score Not benchmarked
  10. 10 Meta Llama 3.1 8B Instruct meta-llama/Meta-Llama-3.1-8B-Instruct 8,514 likes 65 trending score Not benchmarked
  11. 11 Z.ai GLM 5.3 zai-org/GLM-5.3 2,264 likes 58 trending score 73.1coding
  12. 12 Z.ai GLM 5.3 Flash zai-org/GLM-5.3-Flash 2,890 likes 55 trending score 71.0coding
  13. 13 MoonshotAI Kimi K3 moonshotai/Kimi-K3 11.8K likes 49 trending score 61.8coding
  14. 14 Xiaomi MiMo-V2.6-Pro XiaomiMiMo/MiMo-V2.6-Pro-RL 750 likes 44 trending score 72.6coding
  15. 15 Google Gemma 4 31B google/gemma-4-31B-it 4,160 likes 39 trending score 38.3coding
  16. 16 Meta Muse Glimmer 30B meta-models/Muse-Glimmer-30B 2,060 likes 38 trending score Not benchmarked
  17. 17 Qwen3.5-9B Qwen/Qwen3.5-9B 2,206 likes 32 trending score Not benchmarked
  18. 18 Google Gemma 4 26B A4B google/gemma-4-26B-A4B-it 1,679 likes 29 trending score 22.8coding
  19. 19 Qwen3.6 35B A3B Qwen/Qwen3.6-35B-A3B 3,010 likes 29 trending score Not benchmarked
  20. 20 OpenAI gpt-oss-120b openai/gpt-oss-120b 5,426 likes 28 trending score 34.5coding

Most downloaded open-weight models

Downloads over the last 30 days, as Hugging Face counts them, for the open-weight models in our catalog.

  1. 1 Google Gemma 4 26B A4B google/gemma-4-26B-A4B-it 1,679 likes 10.29M downloads in 30 days 22.8coding
  2. 2 Google Gemma 4 31B google/gemma-4-31B-it 4,160 likes 8.97M downloads in 30 days 38.3coding
  3. 3 Qwen3 14B Qwen/Qwen3-14B 510 likes 8.41M downloads in 30 days Not benchmarked
  4. 4 Qwen3.5-9B Qwen/Qwen3.5-9B 2,206 likes 7.48M downloads in 30 days Not benchmarked
  5. 5 Qwen3.8 27B Qwen/Qwen3.8-27B 18K likes 7.46M downloads in 30 days 73.6coding
  6. 6 Qwen2.5 7B Instruct Qwen/Qwen2.5-7B-Instruct 2,446 likes 6.78M downloads in 30 days Not benchmarked
  7. 7 Z.ai GLM 5.3 Flash zai-org/GLM-5.3-Flash 2,890 likes 6.54M downloads in 30 days 71.0coding
  8. 8 OpenAI gpt-oss-20b openai/gpt-oss-20b 5,192 likes 5.9M downloads in 30 days Not benchmarked
  9. 9 Qwen3 VL 8B Instruct Qwen/Qwen3-VL-8B-Instruct 1,214 likes 5.79M downloads in 30 days Not benchmarked
  10. 10 Meta Llama 3.1 8B Instruct meta-llama/Meta-Llama-3.1-8B-Instruct 8,514 likes 5.49M downloads in 30 days Not benchmarked
  11. 11 DeepSeek V3.2 deepseek-ai/DeepSeek-V3.2 1,583 likes 5.29M downloads in 30 days 54.4coding
  12. 12 Meta Llama 3.2 1B Instruct meta-llama/Llama-3.2-1B-Instruct 1,866 likes 5.07M downloads in 30 days Not benchmarked
  13. 13 DeepSeek V4 Flash 0731 deepseek-ai/DeepSeek-V4-Flash-0731 4,083 likes 4.33M downloads in 30 days Not benchmarked
  14. 14 Qwen3.6 35B A3B Qwen/Qwen3.6-35B-A3B 3,010 likes 3.65M downloads in 30 days Not benchmarked
  15. 15 P PrismML Ternary Bonsai 2 27B prism-ml/Ternary-Bonsai-2-27B-gguf 2,758 likes 3.27M downloads in 30 days Not benchmarked
  16. 16 OpenAI gpt-oss-120b openai/gpt-oss-120b 5,426 likes 3.18M downloads in 30 days 34.5coding
  17. 17 Qwen3 32B Qwen/Qwen3-32B 769 likes 2.91M downloads in 30 days 40.0coding
  18. 18 Qwen3.8 Flash Qwen/Qwen3.8-Flash-Next 6,336 likes 2.49M downloads in 30 days 87.4coding
  19. 19 Qwen3.6 27B Qwen/Qwen3.6-27B 2,331 likes 2.31M downloads in 30 days Not benchmarked
  20. 20 Z.ai GLM 5.3 zai-org/GLM-5.3 2,264 likes 1.85M downloads in 30 days 73.1coding

Why closed models are not ranked here

Popularity here is what Hugging Face publishes for open-weight repositories. Closed models such as Claude, GPT and Gemini publish no usage figures, and OpenRouter keeps its usage rankings out of its public API, so we do not rank them on popularity rather than guess.

Best for lists · Coding ranking