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 drifted Gemma 4 31B OpenRouter display price (#14825)

Nicholas Tindle7e44cf7492de
diff --git a/autogpt_platform/backend/backend/data/llm_registry/catalog.py b/autogpt_platform/backend/backend/data/llm_registry/catalog.py
index 5095a46e19..9066cf84b6 100644
--- a/autogpt_platform/backend/backend/data/llm_registry/catalog.py
+++ b/autogpt_platform/backend/backend/data/llm_registry/catalog.py
@@ -641,10 +641,13 @@ def _build_catalog() -> CatalogPayload:
context_window=262144,
max_output_tokens=16384,
price_tier=1,
+ # Live OpenRouter rate as of 2026-09-23: $0.09/$0.34 per 1M,
+ # cache read $0.05/1M (drifted from the $0.14/$0.40 launch price).
cost=CatalogModelCost(
run_credits=1,
- input_credits_per_1m=21.0,
- output_credits_per_1m=60.0,
+ input_credits_per_1m=13.5,
+ output_credits_per_1m=51.0,
+ cache_read_credits_per_1m=7.5,
),
),
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 980da0173b..0f8120646c 100644
--- a/autogpt_platform/backend/backend/data/llm_registry/catalog_test.py
+++ b/autogpt_platform/backend/backend/data/llm_registry/catalog_test.py
@@ -465,15 +465,15 @@ def test_ling_3_0_flash_vl_bills_at_authored_rates():
def test_gemma_4_31b_it_bills_at_authored_rates():
- """Google Gemma 4 31B (OpenRouter, list price $0.14/$0.40 per 1M) —
- flat tier and per-1M projections must match the authored catalog
- entry."""
+ """Google Gemma 4 31B (OpenRouter live rate $0.09/$0.34 per 1M,
+ $0.05/1M cached input as of 2026-09-23) — flat tier and per-1M
+ projections must match the authored catalog entry."""
gemma = LLMModel("google/gemma-4-31b-it")
assert MODEL_COST[gemma] == 1
assert TOKEN_COST[gemma].model_dump() == {
- "input": 21.0,
- "output": 60.0,
- "cache_read": 0.0,
+ "input": 13.5,
+ "output": 51.0,
+ "cache_read": 7.5,
"cache_creation": 0.0,
}
assert MODEL_METADATA[gemma].max_output_tokens == 16384