up to −95%below each provider's official API price. The struck-through price is the official one.

Monthly cost estimator

Coding agents typically re-use 50–80% of their prompt from cache. Adjust to your workload and the last column updates live.

Input tokens / month 20MOutput tokens / month 2MCache hit rate 60%

All 45Anthropic11OpenAI8DeepSeek8Zhipu GLM7xAI Grok3Xiaomi MiMo2Moonshot Kimi2Google Gemini2QWQwen2

FrontierFast & cheapUncensored

Sort: popularSort: biggest savingsSort: cheapestSort: name

Model	Context	Input /1M	Output /1M	Cached /1M	Est. / mo
Claude Fable 5−95%claude-fable-5trial$0.90 in · $0.90 out /1M	1M	$0.90 $10.00	$0.90 $50.00	$0.08	$9.96
GPT-6 Astra−95%gpt-6-astratrial$0.90 in · $0.90 out /1M	1.05M	$0.90 $10.00	$0.90 $50.00	$0.08	$9.96
DeepSeek V4.1 Flash−84%deepseek-v4.1-flashtrial$0.04 in · $0.04 out /1M	1M	$0.04 $0.15	$0.04 $0.60	$0.01	$0.52
GLM-5.2−90%glm-5.2$0.20 in · $0.20 out /1M	1M	$0.20 $1.40	$0.20 $4.40	$0.06	$2.72
Grok 4.7−93%grok-4.7trial$0.20 in · $0.20 out /1M	500K	$0.20 $2.00	$0.20 $6.00	$0.04	$2.48
MiMo V2.6 Promimo-v2.6-pro$0.20 in · $0.20 out /1M	600K	$0.20	$0.20	$0.006	$2.07
Kimi K3kimi-k3trial$0.20 in · $0.20 out /1M	1M	$0.20	$0.20	$0.06	$2.72
Gemma 4 31B Abliteratedgemma-4-31b-abliterated$0.24 in · $0.72 out /1M	33K	$0.24	$0.72	$0.24	$6.24
QWQwen3 Uncensoredqwen3-uncensored$0.49 in · $2.37 out /1M	33K	$0.49	$2.37	$0.49	$14.52
Claude Sonnet 5−77%claude-sonnet-5trial$0.90 in · $0.90 out /1M	1M	$0.90 $2.00	$0.90 $10.00	$0.08	$9.96
GPT-6.1 Sol−95%gpt-6.1-soltrial$0.20 in · $0.20 out /1M	1.05M	$0.20 $2.00	$0.20 $10.00	$0.04	$2.48
DeepSeek V4 Flash−77%deepseek-v4-flash-0731trial$0.06 in · $0.06 out /1M	1M	$0.06 $0.15	$0.06 $0.60	$0.0053	$0.66

Show all 45 models

Estimates assume the cache hit rate applies to input tokens. Actual cost depends on your prompts; see per-request costs in the dashboard. Savings compare our price with the provider's published API price (standard tier, off-peak where the provider has peak pricing) on a blended 3:1 input:output basis.