Model catalog
Pick a model. Send a request.
Frontier open weights, curated. Served on AWS Trainium and NVIDIA Blackwell. One OpenAI-compatible API for every row, and your fine-tunes price the same as the base model.
GPT-OSS
Kimi K2.6Kimi K2.6: Ultra-fastKimi K2.7 Code
Qwen
GLM-5.2Featured
The brand strip. What we'd point at first.
OpenAI
GPT-OSS 120B
Cheapest path to GPT-4-class output. OpenAI's first open-weight model since GPT-2.
Contact us for enterprise capacity and dedicated deployments
Contact salesMoonshot
Kimi K2.6
Coding agents and long-horizon tool use. The model engineering teams reach for when the cheap ones stop being enough.
- Input / 1M
- $0.66
- Cached / 1M
- $0.144
- Output / 1M
- $3.41
- TTFT
- 280MS
- Context
- 256K
Moonshot
Kimi K2.6: Fast
Kimi K2.6 on a low-latency route — the first token comes back faster than the default, in exchange for lower per-route capacity.
- Input / 1M
- $0.68
- Cached / 1M
- $0.144
- Output / 1M
- $3.41
- TTFT
- 280MS
- Context
- 256K
Moonshot
Kimi K2.7 Code
Moonshot's newest flagship MoE — the latest Kimi. Contact us for early-access and reserved capacity.
Contact us for enterprise capacity and dedicated deployments
Contact salesAlibaba
Qwen3 235B
Multilingual coverage and tool use at a price point that holds up against the proprietary mid-tier.
Contact us for enterprise capacity and dedicated deployments
Contact salesZhipu
GLM-5.2
Zhipu's frontier open weight — MLA+DSA MoE with MTP, strong at coding and agentic tool use. Served at NVFP4 on Blackwell.
- Input / 1M
- $1.20
- Cached / 1M
- $0.2
- Output / 1M
- $3.00
- TTFT
- 300MS
- Context
- 1M
All models
Filter by family, search by capability. Every row above is in this table too; this is the scan view.
| Model | Context | Std tok/s | TTFT | $/1M in | $/1M cached | $/1M out | |
|---|---|---|---|---|---|---|---|
GPT-OSS 120B OpenAI · 120B (MoE, ~5B active) · Apache 2.0 | — | — | — | Contact sales | View | ||
Kimi K2.6 Moonshot · 1T MoE (~32B active) · Modified MIT | 256K | 70 | 280ms | $0.66 | $0.144 | $3.41 | View |
Qwen3 235B Alibaba · 235B MoE (~22B active) · Apache 2.0 | — | — | — | Contact sales | View | ||
Kimi K2.6: Fast Moonshot · 1T MoE (~32B active) · Modified MIT | 256K | 70 | 280ms | $0.68 | $0.144 | $3.41 | View |
Kimi K2.7 Code Moonshot · 1T MoE (~32B active) · Modified MIT | — | — | — | Contact sales | View | ||
GLM-5.2 Zhipu · 744B (MoE, ~40B active) · MIT | 1M | 70 | 300ms | $1.20 | $0.2 | $3.00 | View |