AnnexAPI
Claude Haiku 4.5 delivers high-speed responses and efficient execution for everyday language tasks.
0.35 / 1.75After using the "-thinking" suffix with claude-haiku-4-5, forced deep thinking will be enabled.
0.35 / 1.75
GPT5.4 delivers higher quality outputs with fewer iterations through APIs and Codex. It assists individuals and teams in analyzing complex information, building production software, and automating multi-step workflows.
1.25 / 7.50GPT-5.4 deep thinking versions at different levels (low, medium, high, xhigh), with no suffix model having deep thinking set to None.
1.25 / 7.50GPT-5.4 deep thinking versions at different levels (low, medium, high, xhigh), with no suffix model having deep thinking set to None.
1.25 / 7.50GPT-5.4 deep thinking versions at different levels (low, medium, high, xhigh), with no suffix model having deep thinking set to None.
1.25 / 7.50GPT-5.4 deep thinking versions at different levels (low, medium, high, xhigh), with no suffix model having deep thinking set to None.
1.25 / 7.50GPT-5.4 mini enhances the advantages of GPT-5.4 in a faster model.
0.375 / 2.25GPT-5.4 mini enhances the advantages of GPT-5.4 in a faster model. Suffix -low configures reasoning effort to a low level.
0.375 / 2.25GPT-5.4 mini enhances the advantages of GPT-5.4 in a faster model. Suffix -medium configures reasoning effort to a medium level.
0.375 / 2.25GPT-5.4 mini enhances the advantages of GPT-5.4 in a faster model. Suffix -high configures reasoning effort to a high level.
0.375 / 2.25GPT-5.4 mini enhances the advantages of GPT-5.4 in a faster model. Suffix -xhigh configures reasoning effort to an extremely high level.
0.375 / 2.25Claude Sonnet 4.6 balances advanced speed and reasoning capabilities for high-performance production workloads.
1.05 / 5.25Claude Sonnet-4.6's dynamic thinking model automatically decides whether to engage in deep thinking. Suffix -thinking forces standard reasoning.
1.05 / 5.25Claude Sonnet-4.6's dynamic thinking model automatically decides whether to engage in deep thinking. Suffix -low configures reasoning effort to a low level.
1.05 / 5.25Claude Sonnet-4.6's dynamic thinking model automatically decides whether to engage in deep thinking. Suffix -medium configures reasoning effort to a medium level.
1.05 / 5.25Claude Sonnet-4.6's new dynamic thinking model automatically decides whether to engage in deep thinking based on prompts, with suffixes like -max, -high, -medium, and -low determining the thinking level.
1.05 / 5.25Claude Opus 4.6 is Anthropic's powerful frontier model, delivering excellent analysis, reasoning, and detailed generation outputs.
1.75 / 8.75The new Claude-Opus-4-6 dynamic thinking model automatically decides whether to engage in deep thinking. Suffix -thinking forces standard reasoning.
1.75 / 8.75The new Claude-Opus-4-6 dynamic thinking model automatically decides whether to engage in deep thinking. Suffix -low configures reasoning effort to a low level.
1.75 / 8.75The new Claude-Opus-4-6 dynamic thinking model automatically decides whether to engage in deep thinking. Suffix -medium configures reasoning effort to a medium level.
1.75 / 8.75The new Claude-Opus-4-6 dynamic thinking model automatically decides whether to engage in deep thinking based on prompts, with suffixes such as -max, -high, -medium, and -low determining the level of thinking.
1.75 / 8.75The new dynamic thinking model of Claude-Opus-4-8 automatically determines whether deep thinking is required based on prompts.
1.75 / 8.75The new dynamic thinking model of Claude-Opus-4-8 automatically determines whether deep thinking is required based on prompts. Suffix -thinking forces standard reasoning.
1.75 / 8.75The new dynamic thinking model of Claude-Opus-4-8 automatically determines whether deep thinking is required based on prompts. Suffix -low configures reasoning effort to a low level.
1.75 / 8.75The new dynamic thinking model of Claude-Opus-4-8 automatically determines whether deep thinking is required based on prompts. Suffix -medium configures reasoning effort to a medium level.
1.75 / 8.75The new dynamic thinking model of Claude-Opus-4-8 automatically determines whether deep thinking is required based on prompts. Suffixes such as -max, -high, -medium, and -low indicate the level of reasoning.
1.75 / 8.75
Gemini 3.5 Flash delivers cutting-edge intelligence optimized for fast and cost-effective task handling.
1.50 / 9.00Gemini 3.5 Flash configured with real-time web search capabilities for ground-truth information lookups.
1.50 / 9.00Gemini 3.5 Flash with reasoning capabilities enabled for complex multi-step logical tasks.
1.50 / 9.00Gemini 3.5 Flash configured to run without reasoning mode for maximum speed and raw generation rates.
1.50 / 9.00GPT-5.5 is our OpenAI frontier model for the most complex professional work. Reasoning.effort supports: none, low, medium (default), high and xhigh.
2.50 / 15.00
DeepSeek V4 Flash is a hyper-fast, lightweight model optimized for latency-critical tasks and rapid query responses at a flat rate.
0.02 / reqDeepSeek V4 Pro is DeepSeek's advanced reasoning engine, delivering superior coding, mathematics, and complex reasoning capabilities at a flat rate.
0.08 / req
Compared to previous versions, Grok 4.2 has made significant improvements in rapid learning, multi-agent collaboration, multimodal processing, and real-time information acquisition.
0.025 / reqNo description available!
0.025 / req
Qwen3.7 is a next-generation flagship model designed for the era of intelligent agents, with its core strengths lying in the breadth and depth of its agent capabilities: it excels at a wide range of tasks in programming, office work and productivity, as well as long-term autonomous execution.
2.40 / 7.20The high cost-performance Plus model in the Qwen3.7 series has comprehensively upgraded its vision-language capabilities on top of its strong text abilities, while maintaining full agent capabilities in coding, tool usage, and productivity workflows.
0.80 / 3.20
GLM-5.2 is the flagship model for the era of long tasks. It supports a truly usable 1M context, can handle project-level engineering context, executes long-range tasks more stably, follows engineering standards more reliably, and can complete the entire development process from requirements to multi-end deployment in one go.
4.00 / 14.00