AIContenders

Model ranking

A blind comparison of two results for the same task. You learn the model names only after you vote.

This is AIContenders’ internal ranking, not another global model leaderboard. It comes only from comparisons on our portal: blind votes — by users and a clearly labelled judge — on pairs of results of the same task, run by us under the same conditions for every model. So it shows how models do on AIContenders tasks (working pages, animations and games in a single HTML file), not at everything. For a broader, general measure of quality see e.g. Arena (arena.ai) — we use it in the subscription prices.

  1. openai/gpt-6-astra — strength 1.73, 107 games
  2. anthropic/claude-fable-5 — strength 1.68, 139 games
  3. anthropic/claude-opus-5 — strength 1.66, 116 games
  4. openai/gpt-6-luna — strength 1.51, 40 games
  5. meta/muse-spark-1.3 — strength 1.48, 97 games
  6. openai/gpt-5.5 — strength 1.43, 93 games
  7. z-ai/glm-5.3-flash — strength 1.42, 41 games
  8. deepseek/deepseek-v4-flash-0731 — strength 1.42, 82 games
  9. qwen/qwen3.8-max-0902 — strength 1.39, 96 games
  10. moonshotai/kimi-k3 — strength 1.37, 106 games
  11. z-ai/glm-5.3 — strength 1.37, 60 games
  12. openai/gpt-5.6-sol — strength 1.34, 104 games
  13. openai/gpt-6-luna-pro — strength 1.33, 46 games
  14. google/gemini-3.7-flash — strength 1.30, 44 games
  15. anthropic/claude-sonnet-5.5 — strength 1.29, 17 games
  16. unbiased/pareto-26.10 — strength 1.23, 13 games
  17. anthropic/claude-haiku-5.5 — strength 1.23, 16 games
  18. anthropic/claude-sonnet-5 — strength 1.19, 202 games
  19. z-ai/glm-5.3-flashx — strength 1.18, 31 games
  20. openai/gpt-6.1-sol — strength 1.16, 16 games
  21. anthropic/claude-opus-5.5 — strength 1.16, 32 games
  22. qwen/qwen3.8-max-prime — strength 1.16, 8 games
  23. tencent/hy3 — strength 1.09, 48 games
  24. particle-ai/particle-1 — strength 1.08, 7 games
  25. nex-agi/nex-n2.5-pro — strength 1.07, 8 games
  26. qwen/qwen3.8-omni-flash — strength 1.05, 8 games
  27. z-ai/glm-5.3-prime — strength 1.04, 16 games
  28. deepseek/deepseek-v4.1-flash — strength 1.04, 47 games
  29. openai/gpt-6.1-terra-pro — strength 1.03, 9 games
  30. google/gemini-3.5-flash-lite — strength 1.02, 56 games
  31. deepseek/deepseek-v4.1-flash-1017 — strength 1.01, 14 games
  32. google/gemini-3.6-flash — strength 1.00, 41 games
  33. xiaomi/mimo-v2.5 — strength 0.99, 22 games
  34. openai/gpt-6.1-sol-pro — strength 0.98, 16 games
  35. meta/muse-spark-open — strength 0.98, 15 games
  36. x-ai/grok-4.7 — strength 0.97, 8 games
  37. inclusionai/ling-3.1-flash — strength 0.96, 8 games
  38. z-ai/glm-5.3-flashx-legal — strength 0.95, 4 games (few votes yet)
  39. x-ai/grok-4.5 — strength 0.95, 81 games
  40. reflection/beam-1.0-501b-a23b — strength 0.95, 16 games
  41. tencent/hy4 — strength 0.93, 1 games (few votes yet)
  42. google/gemini-3.8-flash — strength 0.92, 11 games
  43. openai/gpt-6.1-terra — strength 0.91, 10 games
  44. minimax/minimax-m3.1-flash — strength 0.90, 15 games
  45. inclusionai/ling-3.0-tiny — strength 0.88, 3 games (few votes yet)
  46. elorian/glimpse — strength 0.88, 12 games
  47. stepfun/step-5-preview — strength 0.87, 16 games
  48. x-ai/grok-4.6 — strength 0.87, 109 games
  49. scaledown/summarize — strength 0.86, 12 games
  50. unbiased/pareto-26.10-preview — strength 0.86, 17 games
  51. xiaomi/mimo-v2.6-pro — strength 0.86, 24 games
  52. deepseek/deepseek-v4-pro-0813 — strength 0.86, 130 games
  53. perceptron/perceptron-mk1.5 — strength 0.83, 16 games
  54. mistralai/mistral-large-4-0 — strength 0.83, 16 games
  55. nvidia/nemotron-3.5-super-vl — strength 0.82, 8 games
  56. nex-agi/nex-n2.5-mini — strength 0.80, 8 games
  57. inception/mercury-3 — strength 0.80, 15 games
  58. deepseek/deepseek-v4-flash — strength 0.79, 8 games
  59. cohere/command-a-plus — strength 0.78, 8 games
  60. prism-ml/ternary-bonsai-2-9b — strength 0.77, 9 games
  61. moondream/moondream3-preview — strength 0.77, 8 games
  62. prism-ml/ternary-bonsai-2-27b — strength 0.76, 16 games
  63. openai/gpt-5.6-terra — strength 0.76, 98 games
  64. xiaomi/mimo-v2.6-flash — strength 0.68, 16 games
  65. qwen/qwen3.8-27b — strength 0.67, 36 games
  66. openai/gpt-5.6-luna — strength 0.61, 108 games
  67. mistralai/mistral-large — strength 0.57, 91 games
  68. poolside/laguna-xs-2.1 — strength 0.55, 40 games
  69. meta-llama/llama-4-scout — strength 0.53, 41 games
  70. qwen/qwen3.7-flash — strength 0.53, 82 games
  71. google/gemini-3.1-pro-preview — strength 0.52, 126 games
  72. meta-llama/llama-4-maverick — strength 0.46, 59 games
  73. anthropic/claude-haiku-4.5 — strength 0.37, 169 games

Based on 1634 blind votes. Model names are revealed only after a vote.