Models
What each model family is for. Benchmarks move monthly and leads are often inside the noise, so these pages focus on context, cost, latency and the tasks each one suits.
| Model | Developer | Context | Input | Output |
|---|---|---|---|---|
| Claude Fable 5.1 | Anthropic | 1M tokens | $10 | $50 |
| Claude Haiku 4.5 | Anthropic | 200K tokens | $1 | $5 |
| Claude Opus 5 | Anthropic | 1M tokens | $5 | $25 |
| Claude Sonnet 5 | Anthropic | 1M tokens | $2 | $10 |
| Command A+ | Cohere | 128K tokens | — | — |
| Composer 2.5 | Cursor (Anysphere) | — | $0.50 | $2.50 |
| DBRX (Discontinued) | Databricks | 32K tokens | — | — |
| DeepSeek V4 | DeepSeek | 1M tokens | $1.32 | $3.96 |
| Doubao Seed 2.0 Code | ByteDance | 256K tokens | — | — |
| ERNIE 5.1 | Baidu | 128K tokens | — | — |
| Gemini 3.8 Flash | 1M tokens | $0.75 | $3.75 | |
| Gemma 4 | Google DeepMind | 128K–256K tokens | — | — |
| GLM-5.3 | Zhipu AI | 1M tokens | $1.40 | $4.40 |
| GPT-6 Astra | OpenAI | 1,050,000 tokens | $10 | $50 |
| Grok 4.6 | xAI | 500K tokens | $2.00 | $6.00 |
| HappyHorse 1.1 | Alibaba Token Hub (ATH) | — | — | — |
| HiDream-O1-Image | HiDream.ai | — | — | — |
| HunyuanImage 3.0 | Tencent | — | — | — |
| Inkling | Thinking Machines Lab | 1M tokens | $1.87 | $4.68 |
| Kimi K3 | Moonshot AI | 1M tokens | $3.00 | $15.00 |
| Kling 3.0 | Kuaishou | — | — | — |
| Ling-2.6-1T | Ant Group (inclusionAI) | 1M tokens native (256K via API) | — | — |
| Llama 4 | Meta | 10M tokens (Scout) | — | — |
| LongCat-2.0 | Meituan | 1M tokens | $0.75 | $2.95 |
| MiMo-V2.5-Pro | Xiaomi | 1M tokens | $0.435 | $0.87 |
| MiniMax H3 (Hailuo 3.0) | MiniMax | — | — | — |
| MiniMax-M3 | MiniMax | 1M tokens | $0.30 | $1.20 |
| Mistral Medium 3.5 | Mistral AI | 256K tokens | $1.50 | $7.50 |
| MOSS-TTS | OpenMOSS | — | — | — |
| Muse Glimmer | Meta | 128K tokens | — | — |
| Muse Spark 1.3 | Meta | 1,048,576 tokens | $1.25 | $4.25 |
| Nano Banana | — | — | — | |
| Nemotron 3 Ultra | NVIDIA | 1M tokens | $0.50 | $2.20 |
| Qwen-Image 3.0 | Alibaba | — | — | — |
| Qwen3.8-Max | Alibaba Cloud | 1M tokens | $2.00 | $6.00 |
| Seed 2.1 (Doubao) | ByteDance | 256K tokens (1M on Doubao-Seed-Evolving) | — | — |
| Seed1.5-VL | ByteDance | — | — | — |
| Seedance 2.5 | ByteDance Seed | — | — | — |
| Seedream 5.0 | ByteDance Seed | — | — | — |
| Sora 2 (Discontinued) | OpenAI | — | — | — |
| Stable Diffusion 3.5 | Stability AI | — | — | — |
| Step-3.7-Flash | StepFun | 256K tokens | $0.20 | $1.15 |
| Tencent Hy4 preview | Tencent | 1M tokens | $0.834 / 1M tokens | $2.501 / 1M tokens |
| Veo 3.1 | Google DeepMind | — | — | — |
| Vidu Q3 Pro | ShengShu Technology | — | — | — |
| Wan 3.0 | Alibaba (Tongyi Lab) | — | — | — |
| Z-Image | Alibaba Tongyi-MAI | — | — | — |
Prices are per million tokens and change several times a year. Every model page carries its own review date — check the vendor before budgeting.
Start with the concepts
Most model comparisons stop making sense without context windows, tokens and quantization. All three are in the glossary.