dh-cli 0.17.1__tar.gz → 0.17.3__tar.gz

{dh_cli-0.17.1 → dh_cli-0.17.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: dh-cli
-Version: 0.17.1
+Version: 0.17.3
 Summary: Dayhoff Labs developer CLI
 Author-email: Dayhoff Labs <dev@dayhofflabs.com>
 License: # PolyForm Noncommercial License 1.0.0

{dh_cli-0.17.1 → dh_cli-0.17.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "dh-cli"
-version = "0.17.1"
+version = "0.17.3"
 description = "Dayhoff Labs developer CLI"
 requires-python = ">=3.11"
 readme = "README.md"

{dh_cli-0.17.1 → dh_cli-0.17.3}/src/dh_cli/bedrock/cost_report.py

@@ -111,6 +111,7 @@ _BASE_MODELS = (
     "opus-4-8",
     "opus-4-7",
     "opus-4-6",
+    "sonnet-5-5",
     "sonnet-5",
     "sonnet-4-6",
     "sonnet-4-5",
@@ -118,6 +119,8 @@ _BASE_MODELS = (
     "fable-5",
     "fable-5-1",
     "gpt-6-astra",
+    "gpt-6-sol",
+    "gpt-6-luna",
     "gpt-5.6-sol",
     "gpt-5.6-terra",
     "gpt-5.6-luna",

{dh_cli-0.17.1 → dh_cli-0.17.3}/src/dh_cli/bedrock/pricing.yaml

@@ -37,6 +37,13 @@ models:
     cache_write_per_1m: 12.50
     # Fable 5.1 cut cache reads 75% vs Fable 5; the only rate that differs.
     cache_read_per_1m: 0.25
+  # Sonnet 5.5: Marketplace rate card `*_global_standard`, 2026-09-29; same
+  # rates as Sonnet 5.
+  sonnet-5-5:
+    input_per_1m: 2.00
+    output_per_1m: 10.00
+    cache_write_per_1m: 2.50
+    cache_read_per_1m: 0.20
   sonnet-5:
     input_per_1m: 2.00
     output_per_1m: 10.00
@@ -65,6 +72,18 @@ models:
     output_per_1m: 50.00
     cache_write_per_1m: 12.50
     cache_read_per_1m: 1.00
+  # GPT-6 Sol/Luna: Marketplace rate card `*_global_standard`, 2026-09-25;
+  # cache write is the 30-minute tier.
+  gpt-6-sol:
+    input_per_1m: 2.00
+    output_per_1m: 10.00
+    cache_write_per_1m: 2.50
+    cache_read_per_1m: 0.20
+  gpt-6-luna:
+    input_per_1m: 0.10
+    output_per_1m: 0.50
+    cache_write_per_1m: 0.125
+    cache_read_per_1m: 0.01
   gpt-5.6-sol:
     input_per_1m: 4.00
     output_per_1m: 20.00

{dh_cli-0.17.1 → dh_cli-0.17.3}/tests/bedrock/test_resolve_base_model.py

@@ -52,6 +52,14 @@ import pytest
         ("anthropic.claude-opus-5-5", "opus-5-5"),
         ("us.anthropic.claude-opus-5-5", "opus-5-5"),
         ("global.anthropic.claude-opus-5-5", "opus-5-5"),
+        # Third occurrence: `sonnet-5` is a substring of `claude-sonnet-5-5`.
+        # Rates are identical, so a collapse leaves totals right but merges the
+        # two models in any per-model breakdown.
+        ("anthropic.claude-sonnet-5", "sonnet-5"),
+        ("us.anthropic.claude-sonnet-5", "sonnet-5"),
+        ("global.anthropic.claude-sonnet-5", "sonnet-5"),
+        ("anthropic.claude-sonnet-5-5", "sonnet-5-5"),
+        ("global.anthropic.claude-sonnet-5-5", "sonnet-5-5"),
         # OpenAI models on `bedrock-runtime` do write invocation logs (only
         # mantle traffic doesn't). Seen in the dev bucket 2026-09, in both the
         # bare and inference-profile-ARN shapes.
@@ -67,6 +75,14 @@ import pytest
             "arn:aws:bedrock:us-east-1:074735440724:inference-profile/global.openai.gpt-6-astra",
             "gpt-6-astra",
         ),
+        ("global.openai.gpt-6-sol", "gpt-6-sol"),
+        ("us.openai.gpt-6-sol", "gpt-6-sol"),
+        ("openai.gpt-6-sol", "gpt-6-sol"),
+        ("global.openai.gpt-6-luna", "gpt-6-luna"),
+        (
+            "arn:aws:bedrock:us-east-1:074735440724:inference-profile/global.openai.gpt-6-luna",
+            "gpt-6-luna",
+        ),
     ],
 )
 def test_resolve_base_model_known(model_id: str, expected_base: str):
@@ -89,6 +105,7 @@ def test_resolve_prefers_the_most_specific_base():
         assert cost_report.resolve_base_model("global.anthropic.claude-fable-5") == "fable-5"
         assert cost_report.resolve_base_model("global.anthropic.claude-opus-5-5") == "opus-5-5"
         assert cost_report.resolve_base_model("global.anthropic.claude-opus-5") == "opus-5"
+        assert cost_report.resolve_base_model("global.anthropic.claude-sonnet-5-5") == "sonnet-5-5"
     finally:
         cost_report._BASE_MODELS = original
 
@@ -132,6 +149,19 @@ def test_opus_55_is_priced_below_opus_5():
         assert pricing["opus-5-5"][rate] < pricing["opus-5"][rate]
 
 
+def test_sonnet_55_rates_match_marketplace_card():
+    """Marketplace `offer-5fu2rhus3byrs`, `*_global_standard`, 2026-09-29."""
+    from dh_cli.bedrock.cost_report import default_pricing_path, load_pricing
+
+    pricing = load_pricing(default_pricing_path())
+    assert pricing["sonnet-5-5"] == {
+        "input_per_1m": 2.00,
+        "output_per_1m": 10.00,
+        "cache_write_per_1m": 2.50,
+        "cache_read_per_1m": 0.20,
+    }
+
+
 @pytest.mark.parametrize(
     "base, rates",
     [
@@ -139,6 +169,10 @@ def test_opus_55_is_priced_below_opus_5():
         # (`USE1-MP:USE1_*_global_standard` unit cost). Sol's regional rate
         # fell from $5.50/$33 in Aug to $4.40/$22 in Sep; Sep is current.
         ("gpt-6-astra", (10.00, 50.00, 12.50, 1.00)),
+        # GPT-6 Sol/Luna: Marketplace rate card, `*_global_standard`,
+        # 2026-09-25 (cache write is the 30-minute tier).
+        ("gpt-6-sol", (2.00, 10.00, 2.50, 0.20)),
+        ("gpt-6-luna", (0.10, 0.50, 0.125, 0.01)),
         ("gpt-5.6-sol", (4.00, 20.00, 5.00, 0.40)),
         ("gpt-5.6-terra", (2.00, 12.00, 2.50, 0.20)),
         ("gpt-5.6-luna", (0.20, 1.20, 0.25, 0.02)),