117B MoE model fine-tuned for Australian context with sovereign data handling, served only from Australian infrastructure, with reasoning, native tool use and a 128k context window
SovereignFastToolsReasoningJSON
- Pricing
- -
High performance coding assistant with reasoning, optimised for algorithms, debugging, and code review, with native tool use and a 192k context, served only from Australian infrastructure
SovereignFastToolsReasoningJSON
- Pricing
- -
Moonshot AI's newest flagship MoE model, extending Kimi-2.7 with stronger reasoning and agentic tool-use performance across a 1M context window
- Input
- US$2.83 / 1M
- Cached input
- US$0.28 / 1M
- Output
- US$14.13 / 1M
Qwen3.8-Max
Alibaba · Preview
Alibaba’s largest and most capable flagship, a 2.4T sparse MoE (95B active) that is natively multimodal over text, image, and video, built for coding, professional work, and long-horizon agentic execution. 1M context
GlobalToolsReasoningJSONVision
17% off
- Input
- US$2.00 / 1MUS$1.65 / 1M
- Cached input
- US$0.25 / 1MUS$0.21 / 1M
- Output
- US$6.00 / 1MUS$4.99 / 1M
Latest flagship model from Z AI, designed for long-horizon agentic tasks with strong reasoning and coding ability. 1M context
55% off
- Input
- US$1.40 / 1MUS$0.61 / 1M
- Cached input
- US$0.26 / 1MUS$0.22 / 1M
- Output
- US$4.40 / 1MUS$1.98 / 1M
MiniMax-M3
MiniMax · Coming soon
428B MoE (roughly 23B active) from MiniMax, combining reasoning, agentic coding and native image and video input in one open-weight model, on a 1M context built on MiniMax Sparse Attention, with roughly a twentieth of the per-token compute of MiniMax M2
GlobalToolsReasoningJSONVision
- Pricing
- Coming soon
Z AI’s flagship, extending GLM-5.2’s long-horizon agentic reasoning and coding across a 1M context, with tool use and a thinking mode
- Input
- US$1.30 / 1M
- Cached input
- US$0.25 / 1M
- Output
- US$4.00 / 1M
Cost-optimised GLM-5.3 tier from Z AI, keeping the 1M context, tool use and thinking mode at roughly a tenth of the flagship rate
- Input
- US$0.13 / 1M
- Cached input
- US$0.02 / 1M
- Output
- US$0.40 / 1M
230B MoE model (10B active, 256 experts) from MiniMax built for software engineering and agentic tool use (56.2% SWE-Bench Pro, matching GPT-5.3-Codex)
SovereignFastToolsReasoningJSON
- Pricing
- -
