LLM Ratings
LLM Leaderboard
Every model scored across standardized benchmarks: OpenRating Index, throughput speed, task cost economics, and input pricing. Sort any column, filter by license, and click a model for the full breakdown.
OpenRating Index vs input price
Input price per 1M tokens (log scale) against the OpenRating Index. Coral marks open-weights models.
Speed Index (Fastest models)
Top ten models ranked by blended output generation speed (tok/s).
Task Economics & Cost Breakdown
Standardized task compute cost broken down into active reasoning (Chain-of-Thought), prompt cache writes, cache hits, output answer, and input tokens.
Cost per Task breaks down empirical spending across standard OpenRating Index benchmark executions. Includes active reasoning tokens, prompt caching writes/hits, and input/output payload fees.
Full Model Ratings
Interactive leaderboard with live sorting and filtering.
| # | Model | License | |||||
|---|---|---|---|---|---|---|---|
| 1 | Claude Opus 5 (max) Anthropic | 63.4 | 59 tok/s | $2.34 | $4.8 | 1.1s | Proprietary |
| 2 | Claude Fable 5 (with fallback) Anthropic | 62.3 | 72 tok/s | $3.14 | $5.5 | 1.0s | Proprietary |
| 3 | GPT-5.6 Sol (max) OpenAI | 61.2 | 70 tok/s | $1.23 | $2.2 | 0.9s | Proprietary |
| 4 | Grok 4.6 (high) xAI | 60.8 | 69 tok/s | $0.84 | $1.8 | 0.8s | Proprietary |
| 5 | Kimi K3 (max) Moonshot AI | 59.8 | 38 tok/s | $0.84 | $0.9 | 1.3s | Open weights |
| 6 | GLM-5.3 (max) Zhipu AI | 59.4 | 95 tok/s | $0.69 | $0.75 | 0.7s | Open weights |
| 7 | Qwen3.8 2.4T A95B Alibaba | 57.9 | 65 tok/s | $1.09 | $0.4 | 0.8s | Open weights |
| 8 | Muse Spark 1.2 (xhigh) Muse AI | 56.9 | 98 tok/s | $0.40 | $0.35 | 0.6s | Proprietary |
| 9 | GPT-5.6 Terra (max) OpenAI | 56.4 | 86 tok/s | $0.51 | $0.6 | 0.7s | Proprietary |
| 10 | Gemini 3.7 Flash (high) | 56.2 | 392 tok/s | $0.40 | $0.3 | 0.4s | Proprietary |
| 11 | DeepSeek V4 Pro 0813 (max) DeepSeek | 53.4 | 82 tok/s | $0.25 | $0.22 | 0.7s | Open weights |
| 12 | GPT-5.6 Luna (max) OpenAI | 52.5 | 149 tok/s | $0.048 | $0.08 | 0.4s | Proprietary |
| 13 | Qwen3.8 27B (xhigh) Alibaba | 52.1 | 129 tok/s | $0.26 | $0.12 | 0.5s | Open weights |
| 14 | Mistral Medium 3.5 Mistral AI | 51.2 | 146 tok/s | $0.46 | $0.25 | 0.5s | Open weights |
| 15 | Claude 4.5 Haiku Anthropic | 49.6 | 195 tok/s | $0.22 | $0.4 | 0.3s | Proprietary |
| 16 | Motif 3 Motif | 47.6 | 111 tok/s | $0.32 | $0.18 | 0.6s | Proprietary |
| 17 | MiniMax-M3 MiniMax | 45.6 | 116 tok/s | $0.14 | $0.1 | 0.5s | Open weights |
| 18 | gpt-oss-120b (high) OpenAI | 44.5 | 164 tok/s | $0.073 | $0.06 | 0.4s | Open weights |
| 19 | Nemotron 3 Super NVIDIA | 43.6 | 124 tok/s | $0.23 | $0.12 | 0.5s | Open weights |
| 20 | Inkling Inkling | 42.5 | 81 tok/s | $0.34 | $0.15 | 0.6s | Proprietary |
| 21 | Nemotron 3 Ultra NVIDIA | 38.6 | 141 tok/s | $0.38 | $0.2 | 0.4s | Open weights |
| 22 | Gemini 3.5 Flash-Lite | 37.6 | 313 tok/s | $0.096 | $0.05 | 0.3s | Proprietary |
| 23 | Solar Open2 250B Upstage | 37.4 | 94 tok/s | $0.19 | $0.14 | 0.6s | Open weights |
| 24 | Nemotron 3.5 Lightning NVIDIA | 36.6 | 285 tok/s | $0.076 | $0.04 | 0.3s | Open weights |
| 25 | Muse Glimmer (high) Muse AI | 35.3 | 229 tok/s | $0.073 | $0.05 | 0.3s | Proprietary |