From 195c2f9e01097f0d420785505a3e697cb49c9db1 Mon Sep 17 00:00:00 2001 From: Bhishmendra Mahala Date: Thu, 29 Oct 2026 10:30:03 +0530 Subject: [PATCH 1/2] feat(bedrock): add ultrafast service tier pricing for GPT-6 Astra Adds execution_modes["ultrafast"] to us.openai.gpt-6-astra, global.openai.gpt-6-astra, openai.gpt-6-astra (bedrock-runtime), and openai.gpt-6-astra (bedrock-mantle). Ultrafast is 6x Standard. Short-context (<=272K) / long-context (>272K) tiers are preserved. Prices per 1M tokens (USD): In-Region / Geo-US: short $66 in / $330 out; long $132 in / $495 out Global CRIS: short $60 in / $300 out; long $120 in / $450 out Source: https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-6-astra.html Co-authored-by: Cursor --- pricing/bedrock-mantle.json | 92 +++++++++++++++++++++++++++ pricing/bedrock.json | 120 ++++++++++++++++++++++++++++++++++++ 2 files changed, 212 insertions(+) diff --git a/pricing/bedrock-mantle.json b/pricing/bedrock-mantle.json index 1e6c3ae3c..500b2dde6 100644 --- a/pricing/bedrock-mantle.json +++ b/pricing/bedrock-mantle.json @@ -974,6 +974,98 @@ } } } + }, + "custom_pricing": { + "context_tier_map": { + "0": "lte-272k", + "272000": "gt-272k" + }, + "regions": { + "default": { + "execution_modes": { + "standard": { + "context_tiers": { + "lte-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0011 + }, + "response_token": { + "price": 0.0055 + }, + "cache_write_input_token": { + "price": 0.001375 + }, + "cache_read_input_token": { + "price": 0.00011 + } + } + } + }, + "gt-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0022 + }, + "response_token": { + "price": 0.00825 + }, + "cache_write_input_token": { + "price": 0.00275 + }, + "cache_read_input_token": { + "price": 0.00022 + } + } + } + } + } + }, + "ultrafast": { + "context_tiers": { + "lte-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0066 + }, + "response_token": { + "price": 0.033 + }, + "cache_write_input_token": { + "price": 0.00825 + }, + "cache_read_input_token": { + "price": 0.00066 + } + } + } + }, + "gt-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0132 + }, + "response_token": { + "price": 0.0495 + }, + "cache_write_input_token": { + "price": 0.0165 + }, + "cache_read_input_token": { + "price": 0.00132 + } + } + } + } + } + } + } + } + } } }, "openai.gpt-6-sol": { diff --git a/pricing/bedrock.json b/pricing/bedrock.json index 7463e2e22..1526d1f6a 100644 --- a/pricing/bedrock.json +++ b/pricing/bedrock.json @@ -17211,6 +17211,46 @@ } } } + }, + "ultrafast": { + "context_tiers": { + "lte-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0066 + }, + "response_token": { + "price": 0.033 + }, + "cache_write_input_token": { + "price": 0.00825 + }, + "cache_read_input_token": { + "price": 0.00066 + } + } + } + }, + "gt-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0132 + }, + "response_token": { + "price": 0.0495 + }, + "cache_write_input_token": { + "price": 0.0165 + }, + "cache_read_input_token": { + "price": 0.00132 + } + } + } + } + } } } } @@ -17281,6 +17321,46 @@ } } } + }, + "ultrafast": { + "context_tiers": { + "lte-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.006 + }, + "response_token": { + "price": 0.03 + }, + "cache_write_input_token": { + "price": 0.0075 + }, + "cache_read_input_token": { + "price": 0.0006 + } + } + } + }, + "gt-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.012 + }, + "response_token": { + "price": 0.045 + }, + "cache_write_input_token": { + "price": 0.015 + }, + "cache_read_input_token": { + "price": 0.0012 + } + } + } + } + } } } } @@ -17421,6 +17501,46 @@ } } } + }, + "ultrafast": { + "context_tiers": { + "lte-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0066 + }, + "response_token": { + "price": 0.033 + }, + "cache_write_input_token": { + "price": 0.00825 + }, + "cache_read_input_token": { + "price": 0.00066 + } + } + } + }, + "gt-272k": { + "pricing_config": { + "pay_as_you_go": { + "request_token": { + "price": 0.0132 + }, + "response_token": { + "price": 0.0495 + }, + "cache_write_input_token": { + "price": 0.0165 + }, + "cache_read_input_token": { + "price": 0.00132 + } + } + } + } + } } } } From c2d25d7c41aa166bf8e4614912f2f36f896d3865 Mon Sep 17 00:00:00 2001 From: Bhishmendra Mahala Date: Fri, 6 Nov 2026 19:04:34 +0530 Subject: [PATCH 2/2] fix(bedrock-mantle/gpt-6-astra): carry additional_units into custom_pricing context tiers The flat pricing_config had additional_units (web_search, file_search) but the custom_pricing context tiers were missing them. Add to all 4 tiers: standard[lte-272k], standard[gt-272k], ultrafast[lte-272k], ultrafast[gt-272k]. --- pricing/bedrock-mantle.json | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) diff --git a/pricing/bedrock-mantle.json b/pricing/bedrock-mantle.json index 500b2dde6..c4b4df006 100644 --- a/pricing/bedrock-mantle.json +++ b/pricing/bedrock-mantle.json @@ -999,6 +999,14 @@ }, "cache_read_input_token": { "price": 0.00011 + }, + "additional_units": { + "web_search": { + "price": 1 + }, + "file_search": { + "price": 0.25 + } } } } @@ -1017,6 +1025,14 @@ }, "cache_read_input_token": { "price": 0.00022 + }, + "additional_units": { + "web_search": { + "price": 1 + }, + "file_search": { + "price": 0.25 + } } } } @@ -1039,6 +1055,14 @@ }, "cache_read_input_token": { "price": 0.00066 + }, + "additional_units": { + "web_search": { + "price": 1 + }, + "file_search": { + "price": 0.25 + } } } } @@ -1057,6 +1081,14 @@ }, "cache_read_input_token": { "price": 0.00132 + }, + "additional_units": { + "web_search": { + "price": 1 + }, + "file_search": { + "price": 0.25 + } } } }