open god mode ▮

Open-weights LLM ranking — overall

# new releases

DeepSeek V4.1 Flash 1017 deepseek-ai · Oct 2026 763.2B · ↑ trending MiniMax M3.1 Flash minimax · Oct 2026 220.5B · ↑ trending Ling 3.1 Flash inclusionai · Oct 2026 572.1B · ↑ trending Beam 1.0 501B A23B reflection · Oct 2026 501.3B · ↑ trending

scored as 70% aa_intelligence · 30% bench_composite (renormalized when signals are missing) — methodology

# model score $/M in ⇄ params ctx prov org
1 deepseek-v4-pro ● 45.6 $0.53 862B 1024k 28 deepseek-ai
2 minimax-m3 ● 36.8 $0.20 427B 1024k 25 minimax
3 k2-horizon-375b-a23b ● 29.6 — 379.2B — — IFM
4 inkling ● 13.3 $0.95 975B 512k 4 thinkingmachines
5 granite-4.2-3b ● 0.7 $0.03 3.7B — 1 ibm-granite
— deepseek-v4.1-flash-1017 ● 100.0 $0.04 763.2B 1024k 45 deepseek-ai
— glm-5.3 ● 92.3 $0.07 753.3B 1024k 67 zai-org
— mimo-v2.6-pro ● 84.6 $0.43 1024.2B 1050k 8 xiaomi
— kimi-k2-0905 ● 81.3 $0.50 1000B (A32B) 256k 3 moonshotai
— kimi-k3 ● 76.9 $0.15 2800B 1024k 42 moonshotai
— minimax-m3.1-flash ● 76.9 $0.05 220.5B 1024k 19 minimax
— minimax-m2.5 ● 72.5 $0.15 229B (A10B) 200k 18 minimax
— hy3 ● 69.8 $0.13 298.8B 256k 9 tencent
— glm-5.3-flash ● 69.2 $0.03 321.3B 1024k 55 zai-org
— qwen3-coder-480b-a35b ● 68.8 $0.25 480B (A35B) 256k 8 qwen
— glm-5 ● 68.7 $0.60 754B 200k 14 zai-org
— mimo-v2.6-flash ● 61.5 $0.09 310.8B 1050k 14 xiaomi
— deepseek-r1 ● 56.3 $0.55 671B (A37B) — 14 deepseek-ai
— kimi-k2.5 ● 56.0 $0.45 1059B (A32B) 256k 11 moonshotai
— muse-spark-open ● 53.8 $1.25 822B 1024k 9 meta-models
— kimi-k2-thinking ● 52.1 $0.30 1000B (A32B) 256k 7 moonshotai
— deepseek-v3.2 ● 47.6 $0.21 671B (A37B) 160k 19 deepseek-ai
— nemotron-3-nano-30b-a3b ● 47.6 $0.05 31.6B 256k 5 nvidia
— kimi-k2.6 ● 45.9 $0.38 1059B (A32B) 256k 29 moonshotai
— beam-1.0-501b-a23b ● 44.8 $0.25 501.3B 1024k 6 reflection
— qwen3-coder-30b-a3b ● 43.8 $0.07 30.5B (A3.3B) 256k 5 qwen
— step-3.5-flash ● 42.3 $0.10 199.4B 256k 3 stepfun-ai
— glm-5.1 ● 40.4 $0.90 754B 200k 21 zai-org
— kimi-k2-instruct ● 40.3 $0.50 1000B (A32B) 128k 5 moonshotai
— deepswe-preview ● 39.6 — 32.8B — — agentica-org
— qwq-32b ● 39.6 $0.15 32.8B — 6 qwen
— glm-5.2 ● 39.3 $0.01 753B 1024k 50 zai-org
— kimi-k2.7-code ● 38.5 $0.38 1059B (A32B) 256k 29 moonshotai
— qwen3.8-27b ● 38.5 $0.01 27.8B 1000k 36 qwen
— qwen3-235b-a22b ● 37.8 $0.20 235B (A22B) — 5 qwen
— glm-4.6 ● 36.0 $0.43 355B (A32B) 200k 8 zai-org
— deepseek-v3 ● 35.4 $0.19 671B (A37B) 160k 14 deepseek-ai
— devstral-small-2512 ● 35.4 — 24B — — mistralai
— qwen3.5-122b-a10b ● 34.1 $0.11 125B (A10B) 256k 7 qwen
— deepseek-v4-flash ● 33.1 $0.05 158B 1024k 32 deepseek-ai
— longcat-flash-thinking-2601 ● 33.1 — 561.9B — — meituan-longcat
— devstral-2-2512 ● 31.3 $0.40 123B 256k 1 mistralai
— llama-4-maverick ● 31.3 $0.12 400B (A17B) 512k 6 meta-llama
— qwen2.5-coder-32b ● 31.3 $0.06 32.8B 32k 6 qwen
— gpt-oss-120b ● 29.2 $0.02 117B (A5.1B) 128k 44 openai
— deepseek-r1-0528 ● 29.0 $0.20 671B (A37B) 160k 10 deepseek-ai
— glm-4.5 ● 28.1 $0.60 355B (A32B) 128k 3 zai-org
— qwen3-vl-235b-a22b-thinking ● 28.0 $0.40 235.7B (A22B) — 4 qwen
— gemma-3-27b-it ● 27.1 $0.08 27B 128k 8 google
— glm-4.7 ● 26.4 $0.40 358B 200k 10 zai-org
— granite-4.2-30b ● 25.6 $0.16 29.3B — 1 ibm-granite
— qwen3.5-27b ● 25.1 $0.20 27.8B 256k 7 qwen
— qwen3.5-35b-a3b ● 25.1 $0.07 36B (A3B) 256k 9 qwen
— qwen3-next-80b-a3b-thinking ● 23.2 $0.14 81.3B (A3B) — 8 qwen
— devstral-small-2505 ● 22.9 — 23.6B — — mistralai
— qwen3-32b ● 21.9 $0.07 32.8B 128k 10 qwen
— qwen3.6-27b ● 20.8 $0.15 27.8B 256k 16 qwen
— step3-vl-10b ● 20.7 — 10.2B — — stepfun-ai
— qwen3.5-397b-a17b ● 20.4 $0.16 403B (A17B) 256k 18 qwen
— k2-horizon-7b ● 19.8 — 9B — — IFM
— minimax-m2.1 ● 19.5 $0.30 229B (A10B) 200k 4 minimax
— glm-4.7-flash ● 16.5 $0.02 31.2B 200k 8 zai-org
— minimax-m2 ● 16.0 $0.15 229B (A10B) 200k 2 minimax
— granite-4.2-8b ● 15.9 $0.06 8.8B 128k 3 ibm-granite
— nemotron-3.5-super-vl ● 15.4 $0.06 — 1024k 12 nvidia
— nemotron-3-super-120b-a12b ● 15.2 $0.08 — 1024k 5 nvidia
— devstral-small-2507 ● 14.6 — 24B — — mistralai
— muse-glimmer-30b ● 13.9 $0.25 29.8B 128k 7 meta-models
— qwen3-vl-30b-a3b-thinking ● 13.4 $0.20 30.5B (A3B) 256k 2 qwen
— qwen3-vl-8b-thinking ● 6.1 — 8.8B — — qwen
— qwen3.6-35b-a3b ● 5.6 $0.02 36B (A3.3B) 256k 21 qwen
— minicpm-sala ● 2.4 — 9.5B — — openbmb
— nemotron-3-ultra-550b-a55b ● 1.9 $0.40 560.5B 256k 8 nvidia
— deepseek-v3.1 ● — $0.25 671B (A37B) 160k 14 deepseek-ai
— gemma-3n-e2b-it ● — — 5.4B — — google
— gemma-4-12b-it ● — $0.30 12B — 2 google
— gemma-4-26b-a4b-it ● — $0.02 26.5B (A4B) 256k 21 google
— gemma-4-31b-it ● — $0.10 32.7B 256k 29 google
— gemma-4-e2b-it ● — — 5.1B — — google
— glm-4.5-air ● — $0.13 106B (A12B) 128k 5 zai-org
— k2-horizon-0.9b ● — — 1.1B — — IFM
— lfm-2.5-2.6b ● — $0.35 2.7B 125k 1 LiquidAI
— llama-3.1-nemotron-nano-8b-v1 ● — — 8B — — nvidia
— longcat-flash-lite ● — — 69.1B — — meituan-longcat
— magistral-small-2506 ● — — 24B — — mistralai
— minimax-m1 ● — $0.40 456B (A45.9B) 1000k 2 minimax
— nvidia-nemotron-nano-9b-v2 ● — $0.04 8.9B — 3 nvidia
— phi-4-reasoning-plus ● — $0.12 14.7B — 1 microsoft
— qwen3-14b ● — $0.07 14.8B 40k 7 qwen
— qwen3-235b-a22b-instruct-2507 ● — $0.09 235.1B (A22B) 256k 11 qwen
— qwen3-30b-a3b ● — $0.10 30.5B (A3.3B) 40k 2 qwen
— qwen3-vl-235b-a22b-instruct ● — $0.20 235.7B (A22B) 256k 10 qwen
— qwen3-vl-30b-a3b-instruct ● — $0.15 30.5B (A3B) 256k 6 qwen
— qwen3-vl-32b-instruct ● — $0.16 33.4B — 5 qwen
— qwen3-vl-4b-instruct ● — — 4.4B — — qwen
— qwen3-vl-8b-instruct ● — $0.08 8.8B 256k 4 qwen
— a.x-k1 ● — — 519B — — skt
— a.x-k2 ● — — 691.7B — — skt
— deepseek-chat-v3-0324 ● — $0.24 684.5B 160k 9 deepseek-ai
— deepseek-v3.1-terminus ● — $0.27 671B (A37B) 160k 3 deepseek-ai
— deepseek-v4-flash-0731 ● — $0.06 158B 1024k 39 deepseek-ai
— deepseek-v4-pro-0813 ● — $0.10 1650.5B 1024k 32 deepseek-ai
— deepseek-v4.1-flash ● — $0.01 763.2B 1024k 55 deepseek-ai
— g9v3-39a5b ● — — 39B — — ai9stars
— g9v3-3b ● — — 3B — — ai9stars
— gemma-2-9b-it ● — — 9.2B — — google
— gemma-3-12b-it ● — $0.05 12B 128k 8 google
— gemma-3n-e4b ● — — 7.8B — — google
— gemma-4-e4b ● — — 8B — — google
— gigachat3.5-432b-a28b ● — — 432B (A28B) — — ai-sage
— glm-4.6v ● — $0.30 107.7B 128k 2 zai-org
— gpt-oss-20b ● — $0.01 21B (A3.6B) 128k 23 openai
— granite-3.3-8b-instruct ● — $0.03 8.2B — 1 ibm-granite
— granite-4.0-h-micro ● — $0.02 3.2B — 1 ibm-granite
— granite-4.1-8b ● — $0.05 8.8B — 1 ibm-granite
— hunyuan-a13b-instruct ● — $0.14 80.4B 128k 1 tencent
— hy-mt2-1.8b ● — $0.04 2B 8k 1 tencent
— hy-mt2-30b-a3b ● — $0.07 30.1B 8k 1 tencent
— hy-mt2-7b ● — $0.07 8B 8k 1 tencent
— hy4-preview ● — $0.75 780B 1024k 11 tencent
— k2-horizon-3.7b ● — — 5.1B — — IFM
— k2-horizon-mova-36b-a4b ● — — 37.4B — — IFM
— k2-think-v2 ● — — 72.6B — — IFM
— k2-v2 ● — — — — — IFM
— kimi-linear-48b-a3b-instruct ● — — 49.1B — — moonshotai
— laguna-m.1 ● — — 225.8B — — poolside
— laguna-s-2.1 ● — $0.09 117.6B — 1 poolside
— laguna-xs-2.1 ● — — 33.4B — — poolside
— laguna-xs.2 ● — — 33.4B — — poolside
— leanstral-1.5-119b-a6b ● — — 119B (A6B) — 1 mistralai
— lfm2-8b-a1b ● — — 8.3B — — LiquidAI
— ling-1t ● — — 1000B (A50B) — — inclusionai
— ling-2.6-1t ● — $0.30 1025.7B — 1 inclusionai
— ling-2.6-flash ● — $0.10 107.5B — 1 inclusionai
— ling-3.0-flash-fin ● — $0.06 127.5B — 2 inclusionai
— ling-3.1-flash ● — $0.30 572.1B 1000k 2 inclusionai
— llama-3.1-405b-instruct ● — $0.12 405B — 4 meta-llama
— llama-3.3-70b-instruct ● — $0.10 70.6B 128k 22 meta-llama
— llama-4-scout ● — $0.09 109B (A17B) 1280k 8 meta-llama
— longcat-2.0 ● — $0.30 1776B 1048k 5 meituan-longcat
— magistral-small-2509 ● — $0.50 24B — 1 mistralai
— mimo-v2.5 ● — $0.12 310.8B 1050k 8 xiaomi
— mimo-v2.5-pro ● — $0.30 1023B 1050k 7 xiaomi
— minicpm5-1b ● — — 1.1B — — openbmb
— minicpm5-2b ● — — 2.5B — — openbmb
— minimax-m2.7 ● — $0.21 229B (A10B) 200k 13 minimax
— ministral-14b-2512 ● — $0.20 13.9B 256k 4 mistralai
— ministral-3-8b-instruct-2512 ● — — 8.9B — — mistralai
— mistral-large-2411 ● — $2.00 123B — 1 mistralai
— mistral-large-3 ● — $0.25 675B (A41B) 256k 4 mistralai
— mistral-large-4-0 ● — $0.70 — 1024k 3 mistralai
— mistral-small-2603 ● — $0.15 119.4B 256k 4 mistralai
— mistral-small-3.2-24b ● — $0.07 24B 250k 5 mistralai
— molmo2-8b ● — — 8.7B — — allenai
— motif-3 ● — $0.50 314.8B — 1 Motif-Technologies
— nemotron-3.5-content-safety ● — $0.05 4.3B 128k 3 nvidia
— nemotron-cascade-2-30b-a3b ● — — 31.6B — — nvidia
— nex-n2-mini ● — — 35.1B — — nex-agi
— nex-n2-pro ● — $0.50 396.8B — 1 nex-agi
— nex-n2.5-mini-free ● — $0.03 35.1B 256k 1 nex-agi
— nex-n2.5-pro-free ● — $0.07 396.8B 256k 1 nex-agi
— olmo-2-32b-instruct ● — — 32B — — allenai
— olmo-3-32b ● — — 32.2B — — allenai
— olmo-3-7b-think ● — — 7.3B — — allenai
— ornith-1.0-35b ● — — 35B — — deepreinforce-ai
— ornith-1.0-397b ● — — 397B — — deepreinforce-ai
— ornith-1.0-9b ● — — 9B — — deepreinforce-ai
— phi-3.5-mini-instruct ● — $0.13 3.8B — 1 microsoft
— phi-4 ● — $0.07 14.7B 16k 4 microsoft
— phi-4-multimodal-instruct ● — $0.08 5.6B — 1 microsoft
— qwen-agentworld-35b-a3b ● — — 34.7B (A3.3B) — — qwen
— qwen2.5-72b-instruct ● — $0.12 72.7B 32k 7 qwen
— qwen2.5-omni-7b ● — — 10.7B — — qwen
— qwen2.5-vl-72b-instruct ● — $0.25 73.4B 125k 5 qwen
— qwen3-30b-a3b-instruct-2507 ● — $0.09 30.5B 256k 5 qwen
— qwen3-coder-next ● — $0.12 79.7B 256k 5 qwen
— qwen3.5-0.8b ● — — 0.9B — — qwen
— qwen3.5-9b ● — $0.02 9.7B 256k 8 qwen
— qwen3.8-2.4t-a95b ● — $2.00 2446.2B 1024k 9 qwen
— qwen3.8-flash-next ● — $0.10 180B — 1 qwen
— ring-2.6-1t ● — $0.30 1025.7B — 1 inclusionai
— solar-open-100b ● — — 102.7B — — upstage
— solar-open2-250b ● — — 250.3B — — upstage
— step-3.7-flash ● — $0.20 201.4B 256k 6 stepfun-ai
— step-5-preview ● — $1.00 604.3B 1024k 7 stepfun-ai
— trinity-large-thinking ● — — — — — arcee-ai
— voxtral-small-24b-2507 ● — $0.10 24.3B 32k 4 mistralai