Skip to content

AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.

fix(llm-registry): correct deepseek-v4.1-flash input price to match live OpenRouter rate (#14820)

Nicholas Tindle9a79ac3d9929
diff --git a/autogpt_platform/backend/backend/data/llm_registry/catalog.py b/autogpt_platform/backend/backend/data/llm_registry/catalog.py
index 502f221030..5095a46e19 100644
--- a/autogpt_platform/backend/backend/data/llm_registry/catalog.py
+++ b/autogpt_platform/backend/backend/data/llm_registry/catalog.py
@@ -504,11 +504,13 @@ def _build_catalog() -> CatalogPayload:
context_window=1048576,
max_output_tokens=384000,
price_tier=1,
+ # Live OpenRouter rate as of 2026-09-23: $0.06/$0.32 per 1M,
+ # cache read $0.01/1M (further drift since this PR's prior fix).
cost=CatalogModelCost(
run_credits=1,
- input_credits_per_1m=22.5,
- output_credits_per_1m=90.0,
- cache_read_credits_per_1m=0.45,
+ input_credits_per_1m=9.0,
+ output_credits_per_1m=48.0,
+ cache_read_credits_per_1m=1.5,
),
),
CatalogModel(
diff --git a/autogpt_platform/backend/backend/data/llm_registry/catalog_test.py b/autogpt_platform/backend/backend/data/llm_registry/catalog_test.py
index ffed154ba8..980da0173b 100644
--- a/autogpt_platform/backend/backend/data/llm_registry/catalog_test.py
+++ b/autogpt_platform/backend/backend/data/llm_registry/catalog_test.py
@@ -340,15 +340,15 @@ def test_qwen3_8_flash_bills_at_authored_rates():
def test_deepseek_v4_1_flash_bills_at_authored_rates():
- """DeepSeek V4.1 Flash (OpenRouter, DeepSeek list price $0.15/$0.60 per
- 1M, $0.003/1M cached input) — flat tier and per-1M projections must
- match the authored catalog entry."""
+ """DeepSeek V4.1 Flash (OpenRouter live rate $0.06/$0.32 per 1M,
+ $0.01/1M cached input as of 2026-09-23) — flat tier and per-1M
+ projections must match the authored catalog entry."""
flash = LLMModel("deepseek/deepseek-v4.1-flash")
assert MODEL_COST[flash] == 1
assert TOKEN_COST[flash].model_dump() == {
- "input": 22.5,
- "output": 90.0,
- "cache_read": 0.45,
+ "input": 9.0,
+ "output": 48.0,
+ "cache_read": 1.5,
"cache_creation": 0.0,
}
assert MODEL_METADATA[flash].max_output_tokens == 384000