Cost-efficiency ranking
How many intelligence points one dollar buys. Higher is better, but remember that models differ in output-token efficiency.
| # | Model | Value | Vs. leader | Composite v5 | Output $ | Type |
|---|---|---|---|---|---|---|
| 1 | Claude Haiku 5.5Anthropic | 181.0 | #1 | 36.2 | $0.50 | |
| 2 | MiniMax M3.1 FlashMiniMax | 110.7 | 61% | 29.0 | $0.60 | |
| 3 | DeepSeek V4.1 Flash Open weightsDeepSeek | 61.9 | 34% | 32.5 | $1.20 | Open weights |
| 4 | MiMo-V2.6 Pro Open weightsXiaomi | 49.0 | 27% | 30.0 | $1.40 | Open weights |
| 5 | GLM-5.3 Open weightsZ.ai | 32.0 | 18% | 32.0 | $2.20 | Open weights |
| 6 | GPT-6.1 TerraOpenAI | 25.5 | 14% | 51.0 | $5.00 | |
| 7 | Gemini 4 CarbonGoogle | 14.8 | 8% | 59.4 | $10.00 | |
| 8 | Claude Sonnet 5.5Anthropic | 13.6 | 8% | 54.2 | $10.00 | |
| 9 | GPT-6.1 SolOpenAI | 13.6 | 8% | 54.5 | $10.00 | |
| 10 | Gemini 4 ArgonGoogle | 12.6 | 7% | 50.4 | $10.00 | |
| 11 | Grok 4.7xAI | 11.7 | 6% | 35.0 | $6.00 | |
| 12 | DeepSeek V4 Pro Open weightsDeepSeek | 11.6 | 6% | 23.0 | $3.96 | Open weights |
| 13 | Kimi K3.1 Pro Open weightsMoonshot AI | 11.5 | 6% | 46.0 | $10.00 | Open weights |
| 14 | Muse WonderMeta | 10.8 | 6% | 43.3 | $10.00 | |
| 15 | Claude Opus 5.5Anthropic | 7.3 | 4% | 58.4 | $20.00 | |
| 16 | GPT-6 AstraOpenAI | 2.8 | 2% | 56.9 | $50.00 | |
| 17 | Claude Fable 5.1Anthropic | 2.5 | 1% | 49.7 | $50.00 |