LLM Leaderboard 2026 - Comparison of AI Models
Comparison and ranking the performance of over 180 AI models (LLMs) across key metrics including quality, price, performance and speed (output speed - tokens per second & latency - TTFT), context window & others. Data sourced from Artificial Analysis. For more details including relating to our methodology, see our FAQs.
| Model | Provider | License | AI Index | GPQA | LiveCode | AIME | Context | Price/1M | Tokens/s | TTFT (s) | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Gemini 3 Pro Preview (high) | Proprietary | 48 | 91% | 92% | 96% | 1m | $4.50 | 126 | 32.20 | ||
| GPT-5.1 (high) | OpenAI | Proprietary | 47 | 87% | 87% | 94% | 400k | $3.44 | 112 | 32.90 | |
| Gemini 3 Flash | Proprietary | 46 | 90% | 91% | 97% | 1m | $1.13 | 224 | 11.79 | ||
| GPT-5.2 (medium) | OpenAI | Proprietary | 45 | 86% | 89% | 97% | 400k | $4.81 | 0 | 0 | |
| Claude Opus 4.5 | Anthropic | Proprietary | 43 | 81% | 74% | 63% | 200k | $10 | 74 | 2.07 | |
| Claude 4.5 Sonnet | Anthropic | Proprietary | 42 | 83% | 71% | 88% | 1m | $6 | 66 | 32.27 | |
| GLM-4.7 | Z AI | Open | 42 | 86% | 89% | 95% | 200k | $0.94 | 117 | 17.63 | |
| GPT-5.1 Codex (high) | OpenAI | Proprietary | 42 | 86% | 85% | 96% | 400k | $3.44 | 144 | 18.17 | |
| Grok 4 | xAI | Proprietary | 41 | 88% | 82% | 93% | 256k | $6 | 43 | 7.47 | |
| DeepSeek V3.2 | DeepSeek | Open | 41 | 84% | 86% | 92% | 128k | $0.32 | 28 | 72.36 | |
| o3 | OpenAI | Proprietary | 41 | 83% | 81% | 88% | 200k | $3.50 | 290 | 11.77 | |
| GPT-5 mini (high) | OpenAI | Proprietary | 41 | 83% | 84% | 91% | 400k | $0.69 | 70 | 114.48 | |
| Gemini 3 Pro Preview (low) | Proprietary | 41 | 89% | 86% | 87% | 1m | $4.50 | 133 | 4.21 | ||
| Kimi K2 Thinking | Kimi | Open | 40 | 84% | 85% | 95% | 256k | $1.07 | 89 | 23.18 | |
| MiniMax-M2.1 | MiniMax | Open | 39 | 83% | 81% | 83% | 205k | $0.53 | 69 | 30.52 | |
| MiMo-V2-Flash | Xiaomi | Open | 39 | 85% | 87% | 96% | 256k | $0.15 | 130 | 16.86 | |
| Grok 4.1 Fast | xAI | Proprietary | 38 | 85% | 82% | 89% | 2m | $0.28 | 171 | 5.50 | |
| GPT-5.1 Codex mini (high) | OpenAI | Proprietary | 38 | 81% | 84% | 92% | 400k | $0.69 | 127 | 10.77 | |
| Claude 4.5 Sonnet | Anthropic | Proprietary | 37 | 73% | 59% | 37% | 1m | $6 | 69 | 2.14 | |
| Claude 4.5 Haiku | Anthropic | Proprietary | 37 | 67% | 62% | 84% | 200k | $2 | 87 | 23.37 | |
| KAT-Coder-Pro V1 | KwaiKAT | Proprietary | 36 | 76% | 75% | 95% | 256k | $0 | 64 | 1.01 | |
| MiniMax-M2 | MiniMax | Open | 36 | 78% | 83% | 78% | 205k | $0.53 | 89 | 23.82 | |
| Nova 2.0 Pro Preview (medium) | Amazon | Proprietary | 35 | 79% | 73% | 89% | 256k | $3.44 | 132 | 34.06 | |
| Doubao-Seed-1.8 | ByteDance Seed | Proprietary | 35 | 80% | 75% | 85% | 256k | $0.15 | 0 | 0 | |
| Gemini 3 Flash | Proprietary | 35 | 81% | 80% | 56% | 1m | $1.13 | 192 | 0.68 | ||
| Grok 4 Fast | xAI | Proprietary | 35 | 85% | 83% | 90% | 2m | $0.28 | 136 | 5.08 | |
| Gemini 2.5 Pro | Proprietary | 34 | 84% | 80% | 88% | 1m | $3.44 | 154 | 35.33 | ||
| DeepSeek V3.2 Speciale | DeepSeek | Open | 34 | 87% | 90% | 97% | 128k | $0.32 | 0 | 0 | |
| GLM-4.7 | Z AI | Open | 34 | 66% | 56% | 48% | 200k | $0.94 | 75 | 0.70 | |
| DeepSeek V3.1 Terminus | DeepSeek | Open | 33 | 79% | 80% | 90% | 128k | $0.80 | 0 | 0 | |
| Doubao Seed Code | ByteDance Seed | Proprietary | 33 | 76% | 77% | 79% | 256k | $0.41 | 0 | 0 | |
| GPT-5.2 | OpenAI | Proprietary | 33 | 71% | 67% | 51% | 400k | $4.81 | 74 | 0.48 | |
| gpt-oss-120B (high) | OpenAI | Open | 33 | 78% | 88% | 93% | 131k | $0.26 | 352 | 6.11 | |
| Qwen3 Max Thinking | Alibaba | Proprietary | 32 | 78% | 54% | 82% | 262k | $2.40 | 35 | 59 | |
| Grok 3 mini Reasoning (high) | xAI | Proprietary | 32 | 79% | 70% | 85% | 1m | $0.35 | 176 | 12.06 | |
| Nova 2.0 Pro Preview (low) | Amazon | Proprietary | 32 | 75% | 64% | 63% | 256k | $3.44 | 135 | 26.93 | |
| DeepSeek V3.2 | DeepSeek | Open | 32 | 75% | 59% | 59% | 128k | $0.32 | 28 | 1.25 | |
| Qwen3 Max | Alibaba | Proprietary | 31 | 76% | 77% | 81% | 262k | $2.40 | 28 | 1.62 | |
| Claude 4.5 Haiku | Anthropic | Proprietary | 30 | 65% | 51% | 39% | 200k | $2 | 108 | 0.58 | |
| Nova 2.0 Lite (medium) | Amazon | Proprietary | 30 | 77% | 66% | 89% | 1m | $0.85 | 250 | 21.82 | |
| Qwen3 235B A22B 2507 | Alibaba | Open | 29 | 79% | 79% | 91% | 256k | $2.63 | 71 | 29.28 | |
| ERNIE 5.0 Thinking Preview | Baidu | Proprietary | 29 | 78% | 81% | 85% | 128k | $1.47 | 0 | 0 | |
| Qwen3 VL 32B | Alibaba | Open | 29 | 73% | 74% | 85% | 256k | $2.63 | 51 | 40.28 | |
| DeepSeek V3.1 Terminus | DeepSeek | Open | 28 | 75% | 53% | 54% | 128k | $0.80 | 0 | 0 | |
| Nova 2.0 Omni (medium) | Amazon | Proprietary | 28 | 76% | 66% | 90% | 1m | $0.85 | 0 | 0 | |
| Kimi K2 0905 | Kimi | Open | 28 | 77% | 61% | 57% | 256k | $1.20 | 60 | 0.55 | |
| Apriel-v1.6-15B-Thinker | ServiceNow | Open | 28 | 73% | 81% | 88% | 128k | $0 | 146 | 13.94 | |
| Qwen3 VL 235B A22B | Alibaba | Open | 27 | 77% | 65% | 88% | 262k | $2.63 | 44 | 46.39 | |
| Magistral Medium 1.2 | Mistral | Proprietary | 27 | 74% | 75% | 82% | 128k | $2.75 | 38 | 53.63 | |
| DeepSeek R1 0528 | DeepSeek | Open | 27 | 81% | 77% | 76% | 128k | $2.36 | 0 | 0 | |
| GPT-5 nano (high) | OpenAI | Proprietary | 27 | 68% | 79% | 84% | 400k | $0.14 | 127 | 148.30 | |
| Qwen3 Next 80B A3B | Alibaba | Open | 27 | 76% | 78% | 84% | 262k | $1.88 | 185 | 11.79 | |
| Grok Code Fast 1 | xAI | Proprietary | 26 | 73% | 66% | 43% | 256k | $0.53 | 234 | 6.73 | |
| Nova 2.0 Lite (low) | Amazon | Proprietary | 25 | 70% | 47% | 47% | 1m | $0.85 | 225 | 15.47 | |
| Qwen3 Coder 480B | Alibaba | Open | 25 | 62% | 59% | 39% | 262k | $3 | 44 | 1.60 | |
| gpt-oss-20B (high) | OpenAI | Open | 25 | 69% | 78% | 89% | 131k | $0.10 | 309 | 6.99 | |
| MiMo-V2-Flash | Xiaomi | Open | 25 | 66% | 40% | 68% | 256k | $0.15 | 106 | 1.41 | |
| NVIDIA Nemotron 3 Nano | NVIDIA | Open | 25 | 76% | 74% | 91% | 1m | $0.10 | 233 | 9.13 | |
| Qwen3 235B 2507 | Alibaba | Open | 24 | 75% | 52% | 72% | 256k | $1.23 | 52 | 1.04 | |
| HyperCLOVA X SEED Think (32B) | Naver | Open | 24 | 62% | 63% | 59% | 128k | $0 | 0 | 0 | |
| Motif-2-12.7B | Motif Technologies | Proprietary | 24 | 70% | 65% | 80% | 128k | $0 | 0 | 0 | |
| Nova 2.0 Omni (low) | Amazon | Proprietary | 24 | 70% | 59% | 56% | 1m | $0.85 | 0 | 0 | |
| gpt-oss-120B (low) | OpenAI | Open | 24 | 67% | 71% | 67% | 131k | $0.26 | 313 | 6.87 | |
| Qwen3 Next 80B A3B | Alibaba | Open | 24 | 74% | 68% | 66% | 262k | $0.88 | 173 | 1.03 | |
| GLM-4.6V | Z AI | Open | 24 | 72% | 16% | 85% | 128k | $0.45 | 74 | 27.63 | |
| Grok 4.1 Fast | xAI | Proprietary | 23 | 64% | 40% | 34% | 2m | $0.28 | 138 | 0.74 | |
| GLM-4.5-Air | Z AI | Open | 23 | 73% | 68% | 81% | 128k | $0.42 | 99 | 20.78 | |
| Nova 2.0 Pro Preview | Amazon | Proprietary | 23 | 64% | 47% | 31% | 256k | $3.44 | 161 | 0.45 | |
| Qwen3 4B 2507 | Alibaba | Open | 23 | 67% | 64% | 83% | 262k | $0 | 0 | 0 | |
| Grok 4 Fast | xAI | Proprietary | 23 | 61% | 40% | 41% | 2m | $0.28 | 135 | 0.58 | |
| Qwen3 30B A3B 2507 | Alibaba | Open | 23 | 71% | 71% | 56% | 262k | $0.75 | 181 | 12.10 | |
| Magistral Small 1.2 | Mistral | Open | 23 | 66% | 72% | 80% | 128k | $0.75 | 204 | 10.14 | |
| Mistral Large 3 | Mistral | Open | 22 | 68% | 47% | 38% | 256k | $0.75 | 49 | 0.61 | |
| EXAONE 4.0 32B | LG AI Research | Open | 22 | 74% | 75% | 80% | 131k | $0.70 | 94 | 21.52 | |
| Gemini 2.5 Flash-Lite (Sep) | Proprietary | 22 | 71% | 69% | 69% | 1m | $0.17 | 599 | 6.27 | ||
| Ring-1T | InclusionAI | Open | 22 | 60% | 64% | 89% | 128k | $0.98 | 0 | 0 | |
| Devstral 2 | Mistral | Open | 22 | 59% | 45% | 37% | 256k | $0 | 51 | 0.44 | |
| Hermes 4 405B | Nous Research | Open | 22 | 73% | 69% | 70% | 128k | $1.50 | 35 | 57.61 | |
| Qwen3 VL 32B | Alibaba | Open | 21 | 67% | 51% | 68% | 256k | $1.23 | 43 | 0.93 | |
| Mistral Medium 3.1 | Mistral | Proprietary | 21 | 59% | 41% | 38% | 128k | $0.80 | 78 | 0.51 | |
| gpt-oss-20B (low) | OpenAI | Open | 21 | 61% | 65% | 62% | 131k | $0.10 | 267 | 8.06 | |
| K2-V2 (high) | MBZUAI Institute of Foundation Models | Open | 21 | 68% | 69% | 78% | 512k | $0 | 0 | 0 | |
| Qwen3 Omni 30B A3B | Alibaba | Open | 21 | 73% | 68% | 74% | 66k | $0.43 | 98 | 21.37 | |
| Qwen3 VL 235B A22B | Alibaba | Open | 21 | 71% | 59% | 71% | 262k | $1.23 | 38 | 1.15 | |
| Ring-flash-2.0 | InclusionAI | Open | 21 | 73% | 63% | 84% | 128k | $0.25 | 85 | 24.87 | |
| Hermes 4 70B | Nous Research | Open | 20 | 70% | 65% | 69% | 128k | $0.20 | 80 | 25.66 | |
| Llama Nemotron Ultra | NVIDIA | Open | 20 | 73% | 64% | 64% | 128k | $0.90 | 36 | 55.71 | |
| Qwen3 Coder 30B A3B | Alibaba | Open | 20 | 52% | 40% | 29% | 262k | $0.90 | 95 | 1.41 | |
| Qwen3 VL 30B A3B | Alibaba | Open | 20 | 70% | 48% | 72% | 256k | $0.35 | 96 | 0.81 | |
| Ling-flash-2.0 | InclusionAI | Open | 20 | 66% | 59% | 65% | 128k | $0.25 | 55 | 1.47 | |
| Gemini 2.5 Flash-Lite (Sep) | Proprietary | 20 | 65% | 64% | 47% | 1m | $0.17 | 387 | 0.40 | ||
| Qwen3 VL 30B A3B | Alibaba | Open | 20 | 72% | 70% | 82% | 256k | $0.75 | 106 | 19.73 | |
| Ling-1T | InclusionAI | Open | 19 | 72% | 68% | 71% | 128k | $0 | 0 | 0 | |
| Devstral Small 2 | Mistral | Open | 19 | 53% | 35% | 34% | 256k | $0 | 197 | 0.37 | |
| Llama Nemotron Super 49B v1.5 | NVIDIA | Open | 19 | 75% | 74% | 77% | 128k | $0.17 | 77 | 26.26 | |
| Nova Premier | Amazon | Proprietary | 19 | 57% | 32% | 17% | 1m | $5 | 79 | 0.82 | |
| K2-V2 (medium) | MBZUAI Institute of Foundation Models | Open | 19 | 60% | 54% | 65% | 512k | $0 | 0 | 0 | |
| Devstral Medium | Mistral | Proprietary | 19 | 49% | 34% | 5% | 256k | $0.80 | 108 | 0.44 | |
| Llama 4 Maverick | Meta | Open | 19 | 67% | 40% | 19% | 1m | $0.42 | 132 | 0.41 | |
| Llama 3.3 Nemotron Super 49B | NVIDIA | Open | 18 | 64% | 28% | 55% | 128k | $0 | 0 | 0 | |
| Nova 2.0 Lite | Amazon | Proprietary | 18 | 60% | 35% | 34% | 1m | $0.85 | 224 | 0.52 | |
| Llama 3.1 405B | Meta | Open | 17 | 52% | 31% | 3% | 128k | $4.19 | 25 | 0.85 | |
| GLM-4.6V | Z AI | Open | 17 | 57% | 41% | 26% | 128k | $0.45 | 48 | 0.90 | |
| ERNIE 4.5 300B A47B | Baidu | Open | 17 | 81% | 47% | 41% | 131k | $0.48 | 24 | 1.95 | |
| Hermes 4 405B | Nous Research | Open | 17 | 54% | 55% | 15% | 128k | $1.50 | 32 | 0.74 | |
| Nova 2.0 Omni | Amazon | Proprietary | 17 | 56% | 31% | 37% | 1m | $0.85 | 219 | 0.69 | |
| Qwen3 VL 8B | Alibaba | Open | 17 | 58% | 35% | 31% | 256k | $0.66 | 63 | 32.78 | |
| OLMo 3 7B Think | Allen Institute for AI | Open | 17 | 52% | 62% | 71% | 66k | $0.14 | 110 | 18.70 | |
| DeepSeek R1 0528 Qwen3 8B | DeepSeek | Open | 16 | 61% | 51% | 64% | 33k | $0.07 | 60 | 34.35 | |
| Ministral 14B (Dec '25) | Mistral | Open | 16 | 57% | 35% | 30% | 256k | $0.20 | 116 | 0.31 | |
| Qwen3 4B 2507 | Alibaba | Open | 16 | 52% | 38% | 52% | 262k | $0 | 0 | 0 | |
| Qwen3 Omni 30B A3B | Alibaba | Open | 16 | 62% | 42% | 52% | 66k | $0.43 | 92 | 0.79 | |
| DeepSeek R1 Distill Llama 70B | DeepSeek | Open | 16 | 40% | 27% | 54% | 128k | $0.88 | 42 | 49.03 | |
| Devstral Small | Mistral | Open | 15 | 41% | 25% | 29% | 256k | $0.15 | 237 | 0.36 | |
| Solar Pro 2 | Upstage | Proprietary | 15 | 69% | 62% | 61% | 66k | $0.50 | 112 | 18.79 | |
| Qwen3 30B A3B 2507 | Alibaba | Open | 15 | 66% | 52% | 66% | 262k | $0.35 | 63 | 0.95 | |
| NVIDIA Nemotron Nano 9B V2 | NVIDIA | Open | 15 | 57% | 72% | 70% | 131k | $0.07 | 112 | 18.05 | |
| Llama Nemotron Super 49B v1.5 | NVIDIA | Open | 15 | 48% | 29% | 8% | 128k | $0.17 | 70 | 0.23 | |
| Ling-mini-2.0 | InclusionAI | Open | 15 | 56% | 43% | 49% | 131k | $0.12 | 176 | 1.44 | |
| K2-V2 (low) | MBZUAI Institute of Foundation Models | Open | 15 | 54% | 39% | 35% | 512k | $0 | 0 | 0 | |
| Mistral Small 3.2 | Mistral | Open | 15 | 51% | 28% | 27% | 128k | $0.15 | 92 | 0.38 | |
| Qwen3 VL 4B | Alibaba | Open | 15 | 49% | 32% | 26% | 256k | $0 | 0 | 0 | |
| Ministral 8B (Dec '25) | Mistral | Open | 15 | 47% | 30% | 32% | 256k | $0.15 | 190 | 0.28 | |
| Llama 3.3 70B | Meta | Open | 15 | 50% | 29% | 8% | 128k | $0.64 | 102 | 0.49 | |
| NVIDIA Nemotron Nano 12B v2 VL | NVIDIA | Open | 15 | 57% | 69% | 75% | 128k | $0.30 | 128 | 15.85 | |
| Qwen3 VL 8B | Alibaba | Open | 15 | 43% | 33% | 27% | 256k | $0.31 | 112 | 0.88 | |
| Llama 3.1 Nemotron Nano 4B v1.1 | NVIDIA | Open | 14 | 41% | 49% | 50% | 128k | $0 | 0 | 0 | |
| Kimi Linear 48B A3B Instruct | Kimi | Open | 14 | 41% | 38% | 36% | 1m | $0 | 0 | 0 | |
| Reka Flash 3 | Reka AI | Open | 14 | 53% | 44% | 34% | 128k | $0.35 | 49 | 42.49 | |
| Llama 3.3 Nemotron Super 49B | NVIDIA | Open | 14 | 52% | 28% | 8% | 128k | $0 | 0 | 0 | |
| Qwen3 VL 4B | Alibaba | Open | 14 | 37% | 29% | 37% | 256k | $0 | 0 | 0 | |
| Solar Pro 2 | Upstage | Proprietary | 14 | 56% | 42% | 30% | 66k | $0.50 | 113 | 0.96 | |
| Llama 4 Scout | Meta | Open | 14 | 59% | 30% | 14% | 10m | $0.28 | 119 | 0.44 | |
| Llama 3.1 Nemotron 70B | NVIDIA | Open | 14 | 47% | 17% | 11% | 128k | $1.20 | 40 | 0.32 | |
| Hermes 4 70B | Nous Research | Open | 14 | 49% | 27% | 11% | 128k | $0.20 | 71 | 0.59 | |
| NVIDIA Nemotron 3 Nano | NVIDIA | Open | 14 | 40% | 36% | 13% | 1m | $0.10 | 228 | 0.54 | |
| Command A | Cohere | Open | 13 | 53% | 29% | 13% | 256k | $4.38 | 97 | 0.21 | |
| NVIDIA Nemotron Nano 9B V2 | NVIDIA | Open | 13 | 56% | 70% | 62% | 131k | $0.10 | 112 | 0.56 | |
| Phi-4 | Microsoft Azure | Open | 13 | 57% | 23% | 18% | 16k | $0.22 | 9 | 0.72 | |
| Qwen3 1.7B | Alibaba | Open | 13 | 36% | 31% | 39% | 32k | $0.40 | 125 | 16.87 | |
| Jamba Reasoning 3B | AI21 Labs | Open | 13 | 33% | 21% | 11% | 262k | $0 | 0 | 0 | |
| R1 1776 | Perplexity | Open | 12 | -- | -- | -- | 128k | $0 | 0 | 0 | |
| Llama 3.2 90B (Vision) | Meta | Open | 12 | 43% | 21% | -- | 128k | $0.72 | 41 | 0.33 | |
| Ministral 3B (Dec '25) | Mistral | Open | 12 | 36% | 25% | 22% | 256k | $0.10 | 269 | 0.27 | |
| EXAONE 4.0 32B | LG AI Research | Open | 12 | 63% | 47% | 39% | 131k | $0.70 | 90 | 0.31 | |
| Nova Micro | Amazon | Proprietary | 12 | 36% | 14% | 6% | 130k | $0.06 | 438 | 0.35 | |
| LFM2 8B A1B | Liquid AI | Open | 11 | 34% | 15% | 25% | 33k | $0 | 0 | 0 | |
| Granite 4.0 H Small | IBM | Open | 11 | 42% | 25% | 14% | 128k | $0.11 | 171 | 8.81 | |
| Granite 4.0 Micro | IBM | Open | 11 | 34% | 18% | 6% | 128k | $0 | 0 | 0 | |
| Phi-4 Mini | Microsoft Azure | Open | 11 | 33% | 13% | 7% | 128k | $0 | 47 | 0.30 | |
| DeepHermes 3 - Mistral 24B | Nous Research | Open | 11 | 38% | 20% | -- | 32k | $0 | 0 | 0 | |
| Llama 3.2 11B (Vision) | Meta | Open | 11 | 22% | 11% | 2% | 128k | $0.16 | 66 | 0.39 | |
| Gemma 3n E4B | Open | 11 | 30% | 15% | 14% | 32k | $0.03 | 46 | 0.35 | ||
| Qwen3 1.7B | Alibaba | Open | 11 | 28% | 13% | 7% | 32k | $0.19 | 120 | 0.80 | |
| Qwen3 0.6B | Alibaba | Open | 11 | 24% | 12% | 18% | 32k | $0.40 | 202 | 10.71 | |
| Gemma 3 27B | Open | 10 | 43% | 14% | 21% | 128k | $0 | 47 | 4.69 | ||
| NVIDIA Nemotron Nano 12B v2 VL | NVIDIA | Open | 10 | 44% | 35% | 27% | 128k | $0.30 | 126 | 0.58 | |
| Phi-4 Multimodal | Microsoft Azure | Open | 10 | 32% | 13% | -- | 128k | $0 | 18 | 0.32 | |
| Gemma 3n E2B | Open | 10 | 23% | 10% | 10% | 32k | $0 | 48 | 0.30 | ||
| Jamba 1.7 Large | AI21 Labs | Open | 9 | 39% | 18% | 2% | 256k | $3.50 | 45 | 0.83 | |
| Molmo 7B-D | Allen Institute for AI | Open | 9 | 24% | 4% | 0% | 4k | $0 | 0 | 0 | |
| Gemma 3 12B | Open | 9 | 35% | 14% | 18% | 128k | $0 | 48 | 8.55 | ||
| Gemma 3 1B | Open | 9 | 24% | 2% | 3% | 32k | $0 | 44 | 0.53 | ||
| Exaone 4.0 1.2B | LG AI Research | Open | 8 | 52% | 52% | 50% | 64k | $0 | 0 | 0 | |
| Gemma 3 270M | Open | 8 | 22% | 0% | 2% | 32k | $0 | 0 | 0 | ||
| Exaone 4.0 1.2B | LG AI Research | Open | 8 | 42% | 29% | 24% | 64k | $0 | 0 | 0 | |
| Granite 4.0 H 1B | IBM | Open | 8 | 26% | 12% | 6% | 128k | $0 | 0 | 0 | |
| OLMo 3 7B | Allen Institute for AI | Open | 8 | 40% | 27% | 41% | 66k | $0.13 | 37 | 0.59 | |
| LFM2 2.6B | Liquid AI | Open | 8 | 31% | 8% | 8% | 33k | $0 | 0 | 0 | |
| DeepHermes 3 - Llama-3.1 8B | Nous Research | Open | 8 | 27% | 9% | -- | 128k | $0 | 0 | 0 | |
| Jamba 1.7 Mini | AI21 Labs | Open | 7 | 32% | 6% | 0% | 258k | $0.25 | 124 | 0.69 | |
| Granite 4.0 1B | IBM | Open | 7 | 28% | 5% | 6% | 128k | $0 | 0 | 0 | |
| Granite 4.0 350M | IBM | Open | 7 | 26% | 2% | 0% | 33k | $0 | 0 | 0 | |
| LFM2 1.2B | Liquid AI | Open | 6 | 23% | 2% | 3% | 33k | $0 | 0 | 0 | |
| Gemma 3 4B | Open | 6 | 29% | 11% | 13% | 128k | $0 | 45 | 0.96 | ||
| Qwen3 0.6B | Alibaba | Open | 6 | 23% | 7% | 10% | 32k | $0.19 | 190 | 0.88 | |
| Granite 4.0 H 350M | IBM | Open | 6 | 26% | 2% | 1% | 33k | $0 | 0 | 0 | |
| DeepSeek-OCR | DeepSeek | Open | -- | -- | -- | -- | 8k | $0.05 | 325 | 0.20 | |
| Grok Voice Agent | xAI | Proprietary | -- | -- | -- | -- | 32k | $0 | 0 | 0 | |
| Olmo 3.1 32B Think | Allen Institute for AI | Open | -- | 59% | 70% | 77% | 66k | $0 | 67 | 30.47 | |
| Cogito v2.1 | Deep Cogito | Open | -- | 77% | 69% | 73% | 128k | $1.25 | 74 | 27.29 | |
| Mi:dm K 2.5 Pro | Korea Telecom | Proprietary | -- | 70% | 66% | 77% | 128k | $0 | 0 | 0 | |
| Mi:dm K 2.5 Pro Preview | Korea Telecom | Proprietary | -- | 72% | 58% | 79% | 128k | $0 | 0 | 0 |
