Price-band ranking
Best AI models $2-10 / 1M tokens
As of August 9, 2026, MoonshotAI: Kimi K3 is the top-quality model priced $2-10 / 1m tokens, scoring 59.7 on the Artificial Analysis Intelligence Index among 18 eligible models in this price tier.
- #1 MoonshotAI: Kimi K3 - Quality 59.7 · $6.00/1M
- #2 Qwen: Qwen3.8 Max - Quality 58.1 · $3.00/1M
- #3 OpenAI: GPT-5.6 Terra (batch) - Quality 56.6 · $2.25/1M
- #4 SpaceXAI: Grok 4.5 - Quality 55.8 · $3.00/1M
- #5 Anthropic: Claude Sonnet 5 (batch) - Quality 55.3 · $4.00/1M
- #6 Meta: Muse Spark 1.1 - Quality 53.2 · $2.00/1M
- #7 OpenAI: GPT-5.4 (batch) - Quality 53.1 · $5.63/1M
- #8 Google: Gemini 3.5 Flash (batch) - Quality 52 · $3.38/1M
| # | Model | Quality | Price/1M |
|---|---|---|---|
| 01 | MoonshotAI: Kimi K3 | 59.7 | $6.00 |
| 02 | Qwen: Qwen3.8 Max | 58.1 | $3.00 |
| 03 | OpenAI: GPT-5.6 Terra (batch) | 56.6 | $2.25 |
| 04 | SpaceXAI: Grok 4.5 | 55.8 | $3.00 |
| 05 | Anthropic: Claude Sonnet 5 (batch) | 55.3 | $4.00 |
| 06 | Meta: Muse Spark 1.1 | 53.2 | $2.00 |
| 07 | OpenAI: GPT-5.4 (batch) | 53.1 | $5.63 |
| 08 | Google: Gemini 3.5 Flash (batch) | 52 | $3.38 |
| 09 | Google: Gemini 3.6 Flash (batch) | 51.6 | $3.00 |
| 10 | Anthropic: Claude Sonnet 4.6 (batch) | 48.4 | $6.00 |
| 11 | Google: Gemini 3.1 Pro Preview (batch) | 47.7 | $4.50 |
| 12 | Qwen: Qwen3.7 Max | 46.7 | $2.21 |
| 13 | OpenAI: GPT-5.1 (batch) | 37.5 | $3.44 |
| 14 | Anthropic: Claude Sonnet 4.5 (batch) | 37.4 | $6.00 |
| 15 | OpenAI: GPT-5 (batch) | 35.3 | $3.44 |
| 16 | Anthropic: Claude Haiku 4.5 (batch) | 29.9 | $2.00 |
| 17 | Anthropic: Claude Sonnet 4 | 29.8 | $6.00 |
| 18 | Google: Gemini 2.5 Pro (batch) | 25.9 | $3.44 |
Source: OpenRouter (openrouter.ai/rankings), as of August 9, 2026.
Benchmark scores: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings).
Methodology: smophy.ai/benchmark/methodology
Token counts originate from each provider's own tokenizer and are not directly comparable across providers.