Sign inSign up

Models

A short, curated list of open-weight models for agentic work — fast first token, high throughput, low prices, and the precision each model runs at shown on its card. We add new ones as they prove out.

Try
glm-5.3flashImageToolsReasoning
TextImageText
Details

Context window: 1M

Max output tokens: 1M

Precision: FP8What does FP8 mean?

10% offuntil Oct 13, 2026, 6:51 PM EDT

Input price: Regular price US$0.15 / M TokensLimited-time price US$0.135 / M Tokens

Output price: Regular price US$0.70 / M TokensLimited-time price US$0.63 / M Tokens

Cache read price: Regular price US$0.03 / M TokensLimited-time price US$0.027 / M Tokens

Reasoning: Low · Medium · High

Latency: 0.5 s

Throughput: 155 tok/s · measured Oct 4, 2026

Uptime 100.0 % · last 3 days

View details

Try
deepseek-v4.1flashToolsReasoning
TextText
Details

Context window: 1M

Max output tokens: 1M

10% offuntil Oct 13, 2026, 6:50 PM EDT

Input price: Regular price US$0.20 / M TokensLimited-time price US$0.18 / M Tokens

Output price: Regular price US$0.89 / M TokensLimited-time price US$0.801 / M Tokens

Cache read price: Regular price US$0.04 / M TokensLimited-time price US$0.036 / M Tokens

Reasoning: Off · Low · Medium · High

Latency: 0.4 s

Throughput: 753 tok/s · measured Oct 10, 2026

Uptime 94.5 % · last 3 days

View details

Try
nemotron-3.530bToolsReasoningSuperFast
TextText
Details

Context window: 256K

Max output tokens: 256K

Precision: BF16What does BF16 mean?

Input price: US$0.07 / M Tokens

Output price: US$0.17 / M Tokens

Cache read price: US$0.02 / M Tokens

Reasoning: Off · On

Latency: 0.3 s

Throughput: 600 tok/s · measured Oct 10, 2026

Uptime 100.0 % · last 3 days

View details

Try
ornith-1.535bImageToolsReasoning
TextImageText
Details

Context window: 256K

Max output tokens: 256K

Precision: FP8What does FP8 mean?

Input price: US$0.14 / M Tokens

Output price: US$0.65 / M Tokens

Cache read price: US$0.025 / M Tokens

Reasoning: Off · On

Latency: 0.3 s

Throughput: 374 tok/s · measured Oct 11, 2026

Uptime 98.2 % · last 3 days

View details

Try
qwen3.827bImageToolsReasoning
TextImageText
Details

Context window: 256K

Max output tokens: 256K

Precision: BF16What does BF16 mean?

Input price: US$0.14 / M Tokens

Output price: US$0.65 / M Tokens

Cache read price: US$0.025 / M Tokens

Reasoning: Off · Low · Medium · High

Latency: 0.4 s

Throughput: 147 tok/s · measured Oct 11, 2026

Uptime 100.0 % · last 12 hours

View details