From 6b27591d5ca833ab01c0c22ac2f3a1faa0e11827 Mon Sep 17 00:00:00 2001 From: Aiden Cline Date: Thu, 10 Sep 2026 20:49:48 -0500 Subject: [PATCH 1/3] fix(amazon-bedrock): correct Meta, Mistral, and Writer catalog --- models/meta/llama-3.3-70b-instruct.toml | 2 +- models/meta/llama-4-scout-17b-instruct.toml | 3 +- models/mistral/magistral-small-2509.toml | 27 +++++++++++++++++ .../ministral-3-14b-instruct-2512.toml | 29 +++++++++++++++++++ .../mistral/ministral-3-3b-instruct-2512.toml | 29 +++++++++++++++++++ .../mistral/ministral-3-8b-instruct-2512.toml | 29 +++++++++++++++++++ models/mistral/mistral-large-2512.toml | 3 +- models/mistral/voxtral-mini-3b-2507.toml | 29 +++++++++++++++++++ models/mistral/voxtral-small-24b-2507.toml | 29 +++++++++++++++++++ models/writer/palmyra-x4.toml | 12 ++++---- models/writer/palmyra-x5.toml | 9 ++++-- .../eu.mistral.pixtral-large-2502-v1:0.toml | 1 + .../meta.llama3-1-70b-instruct-v1:0.toml | 21 ++------------ .../meta.llama3-1-8b-instruct-v1:0.toml | 21 ++------------ .../meta.llama3-3-70b-instruct-v1:0.toml | 21 ++------------ ...eta.llama4-maverick-17b-instruct-v1:0.toml | 22 ++++---------- .../meta.llama4-scout-17b-instruct-v1:0.toml | 23 ++++----------- .../models/mistral.devstral-2-123b.toml | 16 ++-------- .../models/mistral.magistral-small-2509.toml | 18 +++--------- .../mistral.ministral-3-14b-instruct.toml | 18 +++--------- .../mistral.ministral-3-3b-instruct.toml | 18 ++---------- .../mistral.ministral-3-8b-instruct.toml | 19 +++--------- ...mistral.mistral-large-3-675b-instruct.toml | 17 ++--------- .../mistral.pixtral-large-2502-v1:0.toml | 21 ++------------ .../models/mistral.voxtral-mini-3b-2507.toml | 18 ++---------- .../mistral.voxtral-small-24b-2507.toml | 24 ++++----------- .../us.meta.llama3-3-70b-instruct-v1:0.toml | 1 - ...eta.llama4-maverick-17b-instruct-v1:0.toml | 6 ++++ ...s.meta.llama4-scout-17b-instruct-v1:0.toml | 7 +++++ .../us.mistral.pixtral-large-2502-v1:0.toml | 1 + .../models/us.writer.palmyra-x4-v1:0.toml | 9 ++++-- .../models/us.writer.palmyra-x5-v1:0.toml | 14 +++++++-- .../models/writer.palmyra-x4-v1:0.toml | 20 ++++--------- .../models/writer.palmyra-x5-v1:0.toml | 17 ++++------- 34 files changed, 288 insertions(+), 266 deletions(-) create mode 100644 models/mistral/magistral-small-2509.toml create mode 100644 models/mistral/ministral-3-14b-instruct-2512.toml create mode 100644 models/mistral/ministral-3-3b-instruct-2512.toml create mode 100644 models/mistral/ministral-3-8b-instruct-2512.toml create mode 100644 models/mistral/voxtral-mini-3b-2507.toml create mode 100644 models/mistral/voxtral-small-24b-2507.toml diff --git a/models/meta/llama-3.3-70b-instruct.toml b/models/meta/llama-3.3-70b-instruct.toml index 2f48b022bda..246ab54c9ef 100644 --- a/models/meta/llama-3.3-70b-instruct.toml +++ b/models/meta/llama-3.3-70b-instruct.toml @@ -3,7 +3,7 @@ description = "Popular open Llama workhorse for multilingual chat, coding, and s family = "llama" release_date = "2024-12-06" last_updated = "2024-12-06" -attachment = true +attachment = false reasoning = false temperature = true tool_call = true diff --git a/models/meta/llama-4-scout-17b-instruct.toml b/models/meta/llama-4-scout-17b-instruct.toml index e381e973ce9..e129a0ceaac 100644 --- a/models/meta/llama-4-scout-17b-instruct.toml +++ b/models/meta/llama-4-scout-17b-instruct.toml @@ -1,3 +1,4 @@ +# https://github.com/meta-llama/llama-models/blob/main/models/llama4/MODEL_CARD.md name = "Llama 4 Scout 17B Instruct" description = "Open Llama with long-context vision for efficient multimodal agents" family = "llama" @@ -11,7 +12,7 @@ knowledge = "2024-08" open_weights = true [limit] -context = 3_500_000 +context = 10_000_000 output = 16_384 [modalities] diff --git a/models/mistral/magistral-small-2509.toml b/models/mistral/magistral-small-2509.toml new file mode 100644 index 00000000000..f4dcc38c9e9 --- /dev/null +++ b/models/mistral/magistral-small-2509.toml @@ -0,0 +1,27 @@ +# https://docs.mistral.ai/models/magistral-small-1-2-25-09 +# https://huggingface.co/mistralai/Magistral-Small-2509 +# https://huggingface.co/mistralai/Magistral-Small-2509/blob/main/generation_config.json +# Native generation config sets max_new_tokens = 131072, matching the model card's max_tokens setting. +name = "Magistral Small 1.2" +description = "Open multimodal reasoning model for transparent analysis of text and images" +family = "magistral" +release_date = "2025-09-18" +last_updated = "2025-09-18" +attachment = true +reasoning = true +temperature = true +tool_call = true +open_weights = true +license = "Apache 2.0" + +[limit] +context = 131_072 +output = 131_072 + +[modalities] +input = ["text", "image"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mistralai/Magistral-Small-2509" diff --git a/models/mistral/ministral-3-14b-instruct-2512.toml b/models/mistral/ministral-3-14b-instruct-2512.toml new file mode 100644 index 00000000000..f39bfa4efe3 --- /dev/null +++ b/models/mistral/ministral-3-14b-instruct-2512.toml @@ -0,0 +1,29 @@ +# https://docs.mistral.ai/models/ministral-3-14b-25-12 +# https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512 +# https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512/blob/main/generation_config.json +# Native examples request max_tokens = 262144; generation_config.max_length is also 262144. +# This is a shared prompt/generation ceiling, not a host-specific independent output cap. +name = "Ministral 3 14B" +description = "Open vision-language model for efficient local deployment, instruction following, and tool use" +family = "ministral" +release_date = "2025-12-02" +last_updated = "2025-12-02" +attachment = true +reasoning = false +temperature = true +tool_call = true +structured_output = true +open_weights = true +license = "Apache 2.0" + +[limit] +context = 262_144 +output = 262_144 + +[modalities] +input = ["text", "image"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512" diff --git a/models/mistral/ministral-3-3b-instruct-2512.toml b/models/mistral/ministral-3-3b-instruct-2512.toml new file mode 100644 index 00000000000..e3b6fdfeea8 --- /dev/null +++ b/models/mistral/ministral-3-3b-instruct-2512.toml @@ -0,0 +1,29 @@ +# https://mistral.ai/news/mistral-3 +# https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512 +# https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512/blob/main/generation_config.json +# Native examples request max_tokens = 262144; generation_config.max_length is also 262144. +# This is a shared prompt/generation ceiling, not a host-specific independent output cap. +name = "Ministral 3 3B" +description = "Compact open vision-language model for edge deployment, instruction following, and tool use" +family = "ministral" +release_date = "2025-12-02" +last_updated = "2025-12-02" +attachment = true +reasoning = false +temperature = true +tool_call = true +structured_output = true +open_weights = true +license = "Apache 2.0" + +[limit] +context = 262_144 +output = 262_144 + +[modalities] +input = ["text", "image"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512" diff --git a/models/mistral/ministral-3-8b-instruct-2512.toml b/models/mistral/ministral-3-8b-instruct-2512.toml new file mode 100644 index 00000000000..493dc85c43c --- /dev/null +++ b/models/mistral/ministral-3-8b-instruct-2512.toml @@ -0,0 +1,29 @@ +# https://mistral.ai/news/mistral-3 +# https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512 +# https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512/blob/main/generation_config.json +# Native examples request max_tokens = 262144; generation_config.max_length is also 262144. +# This is a shared prompt/generation ceiling, not a host-specific independent output cap. +name = "Ministral 3 8B" +description = "Compact open vision-language model for edge deployment, instruction following, and tool use" +family = "ministral" +release_date = "2025-12-02" +last_updated = "2025-12-02" +attachment = true +reasoning = false +temperature = true +tool_call = true +structured_output = true +open_weights = true +license = "Apache 2.0" + +[limit] +context = 262_144 +output = 262_144 + +[modalities] +input = ["text", "image"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512" diff --git a/models/mistral/mistral-large-2512.toml b/models/mistral/mistral-large-2512.toml index 7efdb8119a5..9905e0e1c5b 100644 --- a/models/mistral/mistral-large-2512.toml +++ b/models/mistral/mistral-large-2512.toml @@ -1,7 +1,8 @@ +# https://mistral.ai/news/mistral-3 name = "Mistral Large 3" description = "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning" family = "mistral-large" -release_date = "2024-11-01" +release_date = "2025-12-02" last_updated = "2025-12-02" attachment = true reasoning = false diff --git a/models/mistral/voxtral-mini-3b-2507.toml b/models/mistral/voxtral-mini-3b-2507.toml new file mode 100644 index 00000000000..af016cdd17d --- /dev/null +++ b/models/mistral/voxtral-mini-3b-2507.toml @@ -0,0 +1,29 @@ +# https://mistral.ai/news/voxtral/ +# https://huggingface.co/mistralai/Voxtral-Mini-3B-2507 +# https://huggingface.co/mistralai/Voxtral-Mini-3B-2507/blob/main/params.json +# https://docs.mistral.ai/api/endpoint/chat +# Native params specify 32768 positions, matching the card's 32k context. +# Output is this shared-window ceiling: prompt + max_tokens must fit; no separate native output cap is published. +name = "Voxtral Mini 3B 2507" +description = "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use" +family = "voxtral" +release_date = "2025-07-15" +last_updated = "2025-07-15" +attachment = true +reasoning = false +temperature = true +tool_call = true +open_weights = true +license = "Apache 2.0" + +[limit] +context = 32_768 +output = 32_768 + +[modalities] +input = ["text", "audio"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mistralai/Voxtral-Mini-3B-2507" diff --git a/models/mistral/voxtral-small-24b-2507.toml b/models/mistral/voxtral-small-24b-2507.toml new file mode 100644 index 00000000000..c42b7c07608 --- /dev/null +++ b/models/mistral/voxtral-small-24b-2507.toml @@ -0,0 +1,29 @@ +# https://mistral.ai/news/voxtral/ +# https://huggingface.co/mistralai/Voxtral-Small-24B-2507 +# https://huggingface.co/mistralai/Voxtral-Small-24B-2507/blob/main/params.json +# https://docs.mistral.ai/api/endpoint/chat +# Native params specify 32768 positions, matching the card's 32k context. +# Output is this shared-window ceiling: prompt + max_tokens must fit; no separate native output cap is published. +name = "Voxtral Small 24B 2507" +description = "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use" +family = "voxtral" +release_date = "2025-07-15" +last_updated = "2025-07-15" +attachment = true +reasoning = false +temperature = true +tool_call = true +open_weights = true +license = "Apache 2.0" + +[limit] +context = 32_768 +output = 32_768 + +[modalities] +input = ["text", "audio"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/mistralai/Voxtral-Small-24B-2507" diff --git a/models/writer/palmyra-x4.toml b/models/writer/palmyra-x4.toml index 265d3ce5855..8721f0a89db 100644 --- a/models/writer/palmyra-x4.toml +++ b/models/writer/palmyra-x4.toml @@ -1,8 +1,10 @@ +# https://writer.com/blog/actions-with-palmyra-x4/ +# https://dev.writer.com/home/models name = "Palmyra X4" -description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use" +description = "Enterprise language model for workflow automation, coding, data analysis, and tool use" family = "palmyra" -release_date = "2025-04-28" -last_updated = "2025-04-28" +release_date = "2024-10-09" +last_updated = "2024-10-09" attachment = false reasoning = true temperature = true @@ -10,8 +12,8 @@ tool_call = true open_weights = false [limit] -context = 122_880 -output = 8_192 +context = 128_000 +output = 4_096 [modalities] input = ["text"] diff --git a/models/writer/palmyra-x5.toml b/models/writer/palmyra-x5.toml index 328c9dc5011..502ff35e2ec 100644 --- a/models/writer/palmyra-x5.toml +++ b/models/writer/palmyra-x5.toml @@ -1,18 +1,21 @@ +# https://writer.com/blog/long-context-palmyra-x5/ +# https://dev.writer.com/home/models +# https://dev.writer.com/home/chat-with-images name = "Palmyra X5" description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use" family = "palmyra" release_date = "2025-04-28" last_updated = "2025-04-28" -attachment = false +attachment = true reasoning = true temperature = true tool_call = true open_weights = false [limit] -context = 1_040_000 +context = 1_000_000 output = 8_192 [modalities] -input = ["text"] +input = ["text", "image"] output = ["text"] diff --git a/providers/amazon-bedrock/models/eu.mistral.pixtral-large-2502-v1:0.toml b/providers/amazon-bedrock/models/eu.mistral.pixtral-large-2502-v1:0.toml index 50089982857..da36a93c809 100644 --- a/providers/amazon-bedrock/models/eu.mistral.pixtral-large-2502-v1:0.toml +++ b/providers/amazon-bedrock/models/eu.mistral.pixtral-large-2502-v1:0.toml @@ -1,3 +1,4 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-pixtral-large.html # EU cross-Region inference profile, verified via ListInferenceProfiles from eu-west-1. # EU geo pricing (Price List eu-west-1 standard): input/output USD per 1M tokens. base_model = "mistral/pixtral-large-2502" diff --git a/providers/amazon-bedrock/models/meta.llama3-1-70b-instruct-v1:0.toml b/providers/amazon-bedrock/models/meta.llama3-1-70b-instruct-v1:0.toml index 24e05e58f65..30c8a50cd36 100644 --- a/providers/amazon-bedrock/models/meta.llama3-1-70b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/meta.llama3-1-70b-instruct-v1:0.toml @@ -1,23 +1,8 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-3-1-70b-instruct.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "meta/llama-3.1-70b-instruct" name = "Llama 3.1 70B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-07-23" -last_updated = "2024-07-23" -attachment = false -reasoning = false -temperature = true -knowledge = "2023-12" -tool_call = true -open_weights = true [cost] input = 0.72 output = 0.72 - -[limit] -context = 128_000 -output = 4_096 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/meta.llama3-1-8b-instruct-v1:0.toml b/providers/amazon-bedrock/models/meta.llama3-1-8b-instruct-v1:0.toml index 0ecb525add1..7d5a8aa6ca7 100644 --- a/providers/amazon-bedrock/models/meta.llama3-1-8b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/meta.llama3-1-8b-instruct-v1:0.toml @@ -1,23 +1,8 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-3-1-8b-instruct.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "meta/llama-3.1-8b-instruct" name = "Llama 3.1 8B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-07-23" -last_updated = "2024-07-23" -attachment = false -reasoning = false -temperature = true -knowledge = "2023-12" -tool_call = true -open_weights = true [cost] input = 0.22 output = 0.22 - -[limit] -context = 128_000 -output = 4_096 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/meta.llama3-3-70b-instruct-v1:0.toml b/providers/amazon-bedrock/models/meta.llama3-3-70b-instruct-v1:0.toml index 80ce307a3e5..49b655c547a 100644 --- a/providers/amazon-bedrock/models/meta.llama3-3-70b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/meta.llama3-3-70b-instruct-v1:0.toml @@ -1,23 +1,8 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-3-3-70b-instruct.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "meta/llama-3.3-70b-instruct" name = "Llama 3.3 70B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2024-12-06" -last_updated = "2024-12-06" -attachment = false -reasoning = false -temperature = true -knowledge = "2023-12" -tool_call = true -open_weights = true [cost] input = 0.72 output = 0.72 - -[limit] -context = 128_000 -output = 4_096 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/meta.llama4-maverick-17b-instruct-v1:0.toml b/providers/amazon-bedrock/models/meta.llama4-maverick-17b-instruct-v1:0.toml index 708c0850c6b..c4c1ab3773d 100644 --- a/providers/amazon-bedrock/models/meta.llama4-maverick-17b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/meta.llama4-maverick-17b-instruct-v1:0.toml @@ -1,23 +1,11 @@ -name = "Llama 4 Maverick 17B Instruct" -description = "Open multimodal Llama model for strong reasoning and fast responses" -family = "llama" -release_date = "2025-04-05" -last_updated = "2025-04-05" -attachment = true -reasoning = false -temperature = true -knowledge = "2024-08" -tool_call = true -open_weights = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-4-maverick-17b-instruct.html +# Converse US profile verified 2026-09-09: maxTokens=8192 succeeds; 16384 exceeds the explicit 8192 limit. +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "meta/llama-4-maverick-17b-instruct" [cost] input = 0.24 output = 0.97 [limit] -context = 1_000_000 -output = 16_384 - -[modalities] -input = ["text", "image"] -output = ["text"] +output = 8_192 diff --git a/providers/amazon-bedrock/models/meta.llama4-scout-17b-instruct-v1:0.toml b/providers/amazon-bedrock/models/meta.llama4-scout-17b-instruct-v1:0.toml index 7f2c62bb743..4dc7fc487db 100644 --- a/providers/amazon-bedrock/models/meta.llama4-scout-17b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/meta.llama4-scout-17b-instruct-v1:0.toml @@ -1,23 +1,12 @@ -name = "Llama 4 Scout 17B Instruct" -description = "Open multimodal Llama model for long-context analysis and efficient agents" -family = "llama" -release_date = "2025-04-05" -last_updated = "2025-04-05" -attachment = true -reasoning = false -temperature = true -knowledge = "2024-08" -tool_call = true -open_weights = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-4-scout-17b-instruct.html +# AWS documents the Bedrock context window as 10M tokens; inherit the matching lab context. +# Converse US profile verified 2026-09-09: maxTokens=8192 succeeds; 16384 exceeds the explicit 8192 limit. +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "meta/llama-4-scout-17b-instruct" [cost] input = 0.17 output = 0.66 [limit] -context = 3_500_000 -output = 16_384 - -[modalities] -input = ["text", "image"] -output = ["text"] +output = 8_192 diff --git a/providers/amazon-bedrock/models/mistral.devstral-2-123b.toml b/providers/amazon-bedrock/models/mistral.devstral-2-123b.toml index 63e02383c7b..a56c3b4be30 100644 --- a/providers/amazon-bedrock/models/mistral.devstral-2-123b.toml +++ b/providers/amazon-bedrock/models/mistral.devstral-2-123b.toml @@ -1,14 +1,8 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-devstral-2-123b.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/devstral-2512" name = "Devstral 2 123B" -description = "Mistral coding agent model for repository tasks and software engineering workflows" -family = "devstral" -release_date = "2026-02-17" -last_updated = "2026-02-17" -attachment = false -reasoning = false -temperature = true -tool_call = true structured_output = true -open_weights = true [cost] input = 0.40 @@ -17,7 +11,3 @@ output = 2.00 [limit] context = 256_000 output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.magistral-small-2509.toml b/providers/amazon-bedrock/models/mistral.magistral-small-2509.toml index 136c31600fd..8e4e75887b1 100644 --- a/providers/amazon-bedrock/models/mistral.magistral-small-2509.toml +++ b/providers/amazon-bedrock/models/mistral.magistral-small-2509.toml @@ -1,15 +1,9 @@ -name = "Magistral Small 1.2" -description = "Mistral reasoning model for transparent analysis, math, and complex decisions" -family = "magistral" -release_date = "2025-12-02" -last_updated = "2025-12-02" -attachment = false -reasoning = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-magistral-small-2509.html +# https://huggingface.co/mistralai/Magistral-Small-2509 — reasoning via the model's think tokens, no effort/budget field. +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/magistral-small-2509" reasoning_options = [] -temperature = true -tool_call = true structured_output = true -open_weights = true [cost] input = 0.50 @@ -18,7 +12,3 @@ output = 1.50 [limit] context = 128_000 output = 40_000 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.ministral-3-14b-instruct.toml b/providers/amazon-bedrock/models/mistral.ministral-3-14b-instruct.toml index c5a71d714e4..bf7801837dc 100644 --- a/providers/amazon-bedrock/models/mistral.ministral-3-14b-instruct.toml +++ b/providers/amazon-bedrock/models/mistral.ministral-3-14b-instruct.toml @@ -1,14 +1,8 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-ministral-14b-3-0.html +# Converse image input verified in us-east-1 on 2026-09-09: a yellow PNG was identified as "Yellow.". +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/ministral-3-14b-instruct-2512" name = "Ministral 14B 3.0" -description = "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads" -family = "ministral" -release_date = "2024-12-01" -last_updated = "2024-12-01" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = false [cost] input = 0.20 @@ -17,7 +11,3 @@ output = 0.20 [limit] context = 128_000 output = 4_096 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.ministral-3-3b-instruct.toml b/providers/amazon-bedrock/models/mistral.ministral-3-3b-instruct.toml index 8905f5dee8a..09cde0e1d1e 100644 --- a/providers/amazon-bedrock/models/mistral.ministral-3-3b-instruct.toml +++ b/providers/amazon-bedrock/models/mistral.ministral-3-3b-instruct.toml @@ -1,14 +1,6 @@ -name = "Ministral 3 3B" -description = "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads" -family = "ministral" -release_date = "2025-12-02" -last_updated = "2025-12-02" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-ministral-3b.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/ministral-3-3b-instruct-2512" [cost] input = 0.10 @@ -17,7 +9,3 @@ output = 0.10 [limit] context = 256_000 output = 8_192 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.ministral-3-8b-instruct.toml b/providers/amazon-bedrock/models/mistral.ministral-3-8b-instruct.toml index e0e28477ebf..3358e8edd66 100644 --- a/providers/amazon-bedrock/models/mistral.ministral-3-8b-instruct.toml +++ b/providers/amazon-bedrock/models/mistral.ministral-3-8b-instruct.toml @@ -1,14 +1,7 @@ -name = "Ministral 3 8B" -description = "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads" -family = "ministral" -release_date = "2024-12-01" -last_updated = "2024-12-01" -attachment = false -reasoning = false -temperature = true -tool_call = true -structured_output = true -open_weights = false +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-ministral-3-8b.html +# Converse image input verified in us-east-1 on 2026-09-09: a blue PNG was identified as "Blue". +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/ministral-3-8b-instruct-2512" [cost] input = 0.15 @@ -17,7 +10,3 @@ output = 0.15 [limit] context = 128_000 output = 4_096 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.mistral-large-3-675b-instruct.toml b/providers/amazon-bedrock/models/mistral.mistral-large-3-675b-instruct.toml index e48e415d27f..ca133c0ad46 100644 --- a/providers/amazon-bedrock/models/mistral.mistral-large-3-675b-instruct.toml +++ b/providers/amazon-bedrock/models/mistral.mistral-large-3-675b-instruct.toml @@ -1,14 +1,7 @@ -name = "Mistral Large 3" -description = "Flagship Mistral model for advanced reasoning, coding, and multilingual work" -family = "mistral" -release_date = "2025-12-02" -last_updated = "2025-12-02" -attachment = false -reasoning = false -temperature = true -tool_call = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-mistral-large-3.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/mistral-large-2512" structured_output = true -open_weights = true [cost] input = 0.50 @@ -17,7 +10,3 @@ output = 1.50 [limit] context = 256_000 output = 8_192 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.pixtral-large-2502-v1:0.toml b/providers/amazon-bedrock/models/mistral.pixtral-large-2502-v1:0.toml index 18325d7d856..20ae752c7b9 100644 --- a/providers/amazon-bedrock/models/mistral.pixtral-large-2502-v1:0.toml +++ b/providers/amazon-bedrock/models/mistral.pixtral-large-2502-v1:0.toml @@ -1,22 +1,7 @@ -name = "Pixtral Large (25.02)" -description = "Mistral vision-language model for image understanding and multimodal chat" -family = "mistral" -release_date = "2025-04-08" -last_updated = "2025-04-08" -attachment = false -reasoning = false -temperature = true -tool_call = true -open_weights = false +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-pixtral-large.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/pixtral-large-2502" [cost] input = 2.00 output = 6.00 - -[limit] -context = 128_000 -output = 8_192 - -[modalities] -input = ["text", "image"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.voxtral-mini-3b-2507.toml b/providers/amazon-bedrock/models/mistral.voxtral-mini-3b-2507.toml index db22b79733b..e6c47fef162 100644 --- a/providers/amazon-bedrock/models/mistral.voxtral-mini-3b-2507.toml +++ b/providers/amazon-bedrock/models/mistral.voxtral-mini-3b-2507.toml @@ -1,23 +1,11 @@ -name = "Voxtral Mini 3B 2507" -description = "Efficient Mistral model for fast chat, extraction, and production assistants" -family = "mistral" -release_date = "2024-12-01" -last_updated = "2024-12-01" -attachment = false -reasoning = false -temperature = true -tool_call = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-voxtral-mini-3b-2507.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/voxtral-mini-3b-2507" structured_output = true -open_weights = false [cost] input = 0.04 output = 0.04 [limit] -context = 128_000 output = 4_096 - -[modalities] -input = ["audio", "text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/mistral.voxtral-small-24b-2507.toml b/providers/amazon-bedrock/models/mistral.voxtral-small-24b-2507.toml index aecea5edbfb..f1ac2ee6791 100644 --- a/providers/amazon-bedrock/models/mistral.voxtral-small-24b-2507.toml +++ b/providers/amazon-bedrock/models/mistral.voxtral-small-24b-2507.toml @@ -1,23 +1,11 @@ -name = "Voxtral Small 24B 2507" -description = "Efficient Mistral model for fast chat, extraction, and production assistants" -family = "mistral" -release_date = "2025-07-01" -last_updated = "2025-07-01" -attachment = true -reasoning = false -temperature = true -tool_call = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-voxtral-small-24b-2507.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +base_model = "mistral/voxtral-small-24b-2507" structured_output = true -open_weights = true [cost] -input = 0.15 -output = 0.35 +input = 0.10 +output = 0.30 [limit] -context = 32_000 -output = 8192 - -[modalities] -input = ["text", "audio"] -output = ["text"] +output = 8_192 diff --git a/providers/amazon-bedrock/models/us.meta.llama3-3-70b-instruct-v1:0.toml b/providers/amazon-bedrock/models/us.meta.llama3-3-70b-instruct-v1:0.toml index c0cdccb6f37..c6596a8c72d 100644 --- a/providers/amazon-bedrock/models/us.meta.llama3-3-70b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/us.meta.llama3-3-70b-instruct-v1:0.toml @@ -2,7 +2,6 @@ # US geo pricing (Price List us-east-1 standard): input/output USD per 1M tokens. base_model = "meta/llama-3.3-70b-instruct" name = "Llama 3.3 70B Instruct (US)" -attachment = false [cost] input = 0.72 diff --git a/providers/amazon-bedrock/models/us.meta.llama4-maverick-17b-instruct-v1:0.toml b/providers/amazon-bedrock/models/us.meta.llama4-maverick-17b-instruct-v1:0.toml index 5030ef4621b..f0d7497b272 100644 --- a/providers/amazon-bedrock/models/us.meta.llama4-maverick-17b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/us.meta.llama4-maverick-17b-instruct-v1:0.toml @@ -1,6 +1,12 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-4-maverick-17b-instruct.html +# Converse verified 2026-09-09: maxTokens=8192 succeeds; 16384 exceeds the explicit 8192 limit. +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. base_model = "meta/llama-4-maverick-17b-instruct" name = "Llama 4 Maverick 17B Instruct (US)" [cost] input = 0.24 output = 0.97 + +[limit] +output = 8_192 diff --git a/providers/amazon-bedrock/models/us.meta.llama4-scout-17b-instruct-v1:0.toml b/providers/amazon-bedrock/models/us.meta.llama4-scout-17b-instruct-v1:0.toml index b0633670585..583f68f426b 100644 --- a/providers/amazon-bedrock/models/us.meta.llama4-scout-17b-instruct-v1:0.toml +++ b/providers/amazon-bedrock/models/us.meta.llama4-scout-17b-instruct-v1:0.toml @@ -1,6 +1,13 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-4-scout-17b-instruct.html +# AWS documents the Bedrock context window as 10M tokens; inherit the matching lab context. +# Converse verified 2026-09-09: maxTokens=8192 succeeds; 16384 exceeds the explicit 8192 limit. +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. base_model = "meta/llama-4-scout-17b-instruct" name = "Llama 4 Scout 17B Instruct (US)" [cost] input = 0.17 output = 0.66 + +[limit] +output = 8_192 diff --git a/providers/amazon-bedrock/models/us.mistral.pixtral-large-2502-v1:0.toml b/providers/amazon-bedrock/models/us.mistral.pixtral-large-2502-v1:0.toml index d0e81bd25a7..15903157777 100644 --- a/providers/amazon-bedrock/models/us.mistral.pixtral-large-2502-v1:0.toml +++ b/providers/amazon-bedrock/models/us.mistral.pixtral-large-2502-v1:0.toml @@ -1,3 +1,4 @@ +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-mistral-ai-pixtral-large.html # US cross-Region inference profile, verified via ListInferenceProfiles from us-east-1. # US geo pricing (Price List us-east-1 standard): input/output USD per 1M tokens. base_model = "mistral/pixtral-large-2502" diff --git a/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml b/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml index f2b5bfb7981..7a3c07d4d4d 100644 --- a/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml +++ b/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml @@ -1,5 +1,6 @@ -# US cross-Region inference profile, verified via ListInferenceProfiles from us-east-1. -# US geo pricing (Price List us-east-1 standard): input/output USD per 1M tokens. +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-palmyra-x4.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +# Bedrock's documented request has no caller-selectable reasoning control. base_model = "writer/palmyra-x4" reasoning_options = [] name = "Palmyra X4 (US)" @@ -7,3 +8,7 @@ name = "Palmyra X4 (US)" [cost] input = 2.50 output = 10.00 + +[limit] +input = 122_880 +output = 8_192 diff --git a/providers/amazon-bedrock/models/us.writer.palmyra-x5-v1:0.toml b/providers/amazon-bedrock/models/us.writer.palmyra-x5-v1:0.toml index b56d9601c9f..0be9e353cb1 100644 --- a/providers/amazon-bedrock/models/us.writer.palmyra-x5-v1:0.toml +++ b/providers/amazon-bedrock/models/us.writer.palmyra-x5-v1:0.toml @@ -1,9 +1,19 @@ -# US cross-Region inference profile, verified via ListInferenceProfiles from us-east-1. -# US geo pricing (Price List us-east-1 standard): input/output USD per 1M tokens. +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-palmyra-x5.html +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-writer-palmyra-x5.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +# Bedrock accepts text only and documents no caller-selectable reasoning control. base_model = "writer/palmyra-x5" +attachment = false reasoning_options = [] name = "Palmyra X5 (US)" [cost] input = 0.60 output = 6.00 + +[limit] +context = 1_040_000 +input = 1_040_000 + +[modalities] +input = ["text"] diff --git a/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml b/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml index d532f6fab3f..797082c4ecb 100644 --- a/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml +++ b/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml @@ -1,23 +1,13 @@ -name = "Palmyra X4" -description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use" -family = "palmyra" -release_date = "2025-04-28" -last_updated = "2025-04-28" -attachment = false -reasoning = true +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-palmyra-x4.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +# Bedrock's documented request has no caller-selectable reasoning control. +base_model = "writer/palmyra-x4" reasoning_options = [] -temperature = true -tool_call = true -open_weights = false [cost] input = 2.5 output = 10 [limit] -context = 122_880 +input = 122_880 output = 8_192 - -[modalities] -input = ["text"] -output = ["text"] diff --git a/providers/amazon-bedrock/models/writer.palmyra-x5-v1:0.toml b/providers/amazon-bedrock/models/writer.palmyra-x5-v1:0.toml index db7bae7cbcc..51974afff22 100644 --- a/providers/amazon-bedrock/models/writer.palmyra-x5-v1:0.toml +++ b/providers/amazon-bedrock/models/writer.palmyra-x5-v1:0.toml @@ -1,14 +1,10 @@ -name = "Palmyra X5" -description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use" -family = "palmyra" -release_date = "2025-04-28" -last_updated = "2025-04-28" +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-palmyra-x5.html +# https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-writer-palmyra-x5.html +# https://aws.amazon.com/bedrock/pricing/ — us-east-1 standard, USD per 1M tokens. +# Bedrock accepts text only and documents no caller-selectable reasoning control. +base_model = "writer/palmyra-x5" attachment = false -reasoning = true reasoning_options = [] -temperature = true -tool_call = true -open_weights = false [cost] input = 0.6 @@ -16,8 +12,7 @@ output = 6 [limit] context = 1_040_000 -output = 8_192 +input = 1_040_000 [modalities] input = ["text"] -output = ["text"] From f91f6c108af68ca96f6d388bedb4e54173908789 Mon Sep 17 00:00:00 2001 From: Aiden Cline Date: Thu, 10 Sep 2026 21:03:23 -0500 Subject: [PATCH 2/3] fix(amazon-bedrock): preserve Palmyra X4 context --- providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml | 2 +- providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml b/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml index 7a3c07d4d4d..98c43dce244 100644 --- a/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml +++ b/providers/amazon-bedrock/models/us.writer.palmyra-x4-v1:0.toml @@ -10,5 +10,5 @@ input = 2.50 output = 10.00 [limit] -input = 122_880 +context = 122_880 output = 8_192 diff --git a/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml b/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml index 797082c4ecb..faf256a92d1 100644 --- a/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml +++ b/providers/amazon-bedrock/models/writer.palmyra-x4-v1:0.toml @@ -9,5 +9,5 @@ input = 2.5 output = 10 [limit] -input = 122_880 +context = 122_880 output = 8_192 From 2ed6123915e95ee2970885b6f31f7795900bc48f Mon Sep 17 00:00:00 2001 From: opencode-agent Date: Thu, 10 Sep 2026 22:33:25 -0500 Subject: [PATCH 3/3] fix(writer): preserve Palmyra update date --- models/writer/palmyra-x4.toml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/models/writer/palmyra-x4.toml b/models/writer/palmyra-x4.toml index 8721f0a89db..2e4fc6728c4 100644 --- a/models/writer/palmyra-x4.toml +++ b/models/writer/palmyra-x4.toml @@ -4,7 +4,7 @@ name = "Palmyra X4" description = "Enterprise language model for workflow automation, coding, data analysis, and tool use" family = "palmyra" release_date = "2024-10-09" -last_updated = "2024-10-09" +last_updated = "2025-04-28" attachment = false reasoning = true temperature = true