Tokens Boutique

Compare and calculate the latest AI model token prices.

fair • transparent • no bs

Cheapest: solar-pro4·$0.003480

100K in/4K out
128 models

Saved workloads

No saved workloads yet.

VendorModelIn $/MOut $/MEst. costContextMax outHQReleased
Upstagesolar-pro4$0.03$0.12$0.003480512K128KKR2026-08-11
Amazonnova-micro$0.04$0.14$0.004060128K5KUS2024-12-05
Amazonnova-lite$0.06$0.24$0.006960300K5KUS2024-12-05
Alibabaqwen3.7-flash$0.03$0.11$0.0096201M64KCN2026-07-15
Mistralministral-3b-latest$0.10$0.10$0.01256K256K*EU2025-12-02
Metamuse-spark-1.2-contributor$0.10$0.20$0.011M128KUS2026-08-05
Googlegemini-2.5-flash-lite$0.10$0.40$0.011M64K
Δ8K
US2025-07-22
Baiduernie-4.5-turbo$0.11$0.45$0.01128K*8K*CN2025-04-25
Xiaomimimo-v2.5$0.14$0.28$0.021M128KCN2026-04-22
Mistralministral-8b-latest$0.15$0.15$0.02256K256K*EU2025-12-02
Tencenthunyuan-hy3$0.14$0.56$0.02256K256K*CN2026-07-06
Mistralmistral-small-latest$0.15$0.60$0.02256K256K*EU2026-03-16
Alibabaqwen3.8-flash$0.16$0.47$0.021M64KCN2026-08-26
Mistralministral-14b-latest$0.20$0.20$0.02256K256K*EU2025-12-02
AI21 Labsjamba-mini$0.20$0.40$0.02256K256K*IL2026-01-08
StepFunstep-3.7-flash$0.20$1.15$0.02256K256K*CN2026-05-29
DeepSeekdeepseek-v4-flash$0.22$0.66$0.021M384KCN2026-04-24
DeepSeekdeepseek-v4-flash-vision-exp$0.22$0.66$0.021M384KCN2026-08-21
OpenAIgpt-5.6-luna$0.20$1.20$0.021.05M128KUS2026-07-09
OpenAIgpt-5.4-nano$0.20$1.25$0.03400K128KUS2026-03-17
DeepSeekdeepseek-v3.2-exp †$0.28$0.42$0.03128K8KCN2025-09-29
DeepSeekdeepseek-v3.2-exp-reasoner †$0.28$0.42$0.03128K64KCN2025-09-29
Alibabaqwen3.6-flash$0.25$1.50$0.031M64KCN2026-04-16
Googlegemini-3.1-flash-lite$0.25$1.50$0.031M64K
Δ8K
US2026-05-07
SpaceXAIgrok-3-mini †$0.30$0.50$0.03128K128K*US2025-02-19
Showing 1–25 of 128 modelsMore load as you scroll
*Max output derived from context size: the vendor sets no separate max output limit, only requires prompt tokens plus output tokens to stay within the context window.
**A real vendor-published limit that applies API-wide rather than being stated per model.
ΔThe API's default output when you don't explicitly request more. The number above it is still the true ceiling.
Retired model. No longer available via the vendor's API.
Deprecated model.
No published cache-hit pricing for this model, showing the regular price instead.
Includes a per-request search fee, already included into Est. cost. (Hover bullet for details)
Separate citation/reasoning/search-query token charges apply, not included in Est. cost. (Hover bullet for details)
Has a prompt cache-write price, a one-time per-session cost not included in Est. cost. (Hover bullet for details)
Model-specific access note. (Hover bullet for details)