Skip to content

Drop in a screenshot and convert it to clean code (HTML/Tailwind/React/Vue)

Python

Remove unused model options

Abi Raja6eff104730a7
diff --git a/backend/agent/providers/anthropic/provider.py b/backend/agent/providers/anthropic/provider.py
index 56271973..e3937453 100644
--- a/backend/agent/providers/anthropic/provider.py
+++ b/backend/agent/providers/anthropic/provider.py
@@ -21,10 +21,7 @@ from agent.providers.token_usage import TokenUsage
from agent.tools import CanonicalToolDefinition, ToolCall, parse_json_arguments
from llm import Llm
-THINKING_MODELS = {
- Llm.CLAUDE_4_5_SONNET_2025_09_29.value,
- Llm.CLAUDE_4_5_OPUS_2025_11_01.value,
-}
+THINKING_MODELS: set[str] = set()
ADAPTIVE_THINKING_MODELS = {
Llm.CLAUDE_OPUS_4_6.value,
Llm.CLAUDE_OPUS_4_8_LOW.value,
diff --git a/backend/agent/providers/pricing.py b/backend/agent/providers/pricing.py
index 5f4e2d5d..e33ba226 100644
--- a/backend/agent/providers/pricing.py
+++ b/backend/agent/providers/pricing.py
@@ -15,15 +15,9 @@ class ModelPricing:
# Pricing keyed by the API model name string sent to the provider.
MODEL_PRICING: Dict[str, ModelPricing] = {
# --- OpenAI ---
- "gpt-4.1-2025-04-14": ModelPricing(
- input=2.00, output=8.00, cache_read=0.50
- ),
"gpt-5.2-codex": ModelPricing(
input=1.75, output=14.00, cache_read=0.4375
),
- "gpt-5.3-codex": ModelPricing(
- input=1.75, output=14.00, cache_read=0.4375
- ),
"gpt-5.4-mini": ModelPricing(
input=0.40, output=3.20, cache_read=0.10
),
@@ -37,12 +31,6 @@ MODEL_PRICING: Dict[str, ModelPricing] = {
"claude-sonnet-4-6": ModelPricing(
input=3.00, output=15.00, cache_read=0.30, cache_write=3.75
),
- "claude-sonnet-4-5-20250929": ModelPricing(
- input=3.00, output=15.00, cache_read=0.30, cache_write=3.75
- ),
- "claude-opus-4-5-20251101": ModelPricing(
- input=5.00, output=25.00, cache_read=0.50, cache_write=6.25
- ),
"claude-opus-4-6": ModelPricing(
input=5.00, output=25.00, cache_read=0.50, cache_write=6.25
),
diff --git a/backend/evals/runner.py b/backend/evals/runner.py
index df43fc8b..40acd8a3 100644
--- a/backend/evals/runner.py
+++ b/backend/evals/runner.py
@@ -166,7 +166,7 @@ async def run_image_evals(
if data_url is None:
data_url = await image_to_data_url(filepath)
current_model_for_task = (
- selected_model if n_idx == 0 else Llm.GPT_4_1_2025_04_14
+ selected_model if n_idx == 0 else Llm.GPT_5_2_CODEX_LOW
)
coro = generate_code_and_time(
image_url=data_url,
diff --git a/backend/llm.py b/backend/llm.py
index 2ab168f6..feb826fc 100644
--- a/backend/llm.py
+++ b/backend/llm.py
@@ -5,15 +5,10 @@ from typing import TypedDict
# Actual model versions that are passed to the LLMs and stored in our logs
class Llm(Enum):
# GPT
- GPT_4_1_2025_04_14 = "gpt-4.1-2025-04-14"
GPT_5_2_CODEX_LOW = "gpt-5.2-codex (low thinking)"
GPT_5_2_CODEX_MEDIUM = "gpt-5.2-codex (medium thinking)"
GPT_5_2_CODEX_HIGH = "gpt-5.2-codex (high thinking)"
GPT_5_2_CODEX_XHIGH = "gpt-5.2-codex (xhigh thinking)"
- GPT_5_3_CODEX_LOW = "gpt-5.3-codex (low thinking)"
- GPT_5_3_CODEX_MEDIUM = "gpt-5.3-codex (medium thinking)"
- GPT_5_3_CODEX_HIGH = "gpt-5.3-codex (high thinking)"
- GPT_5_3_CODEX_XHIGH = "gpt-5.3-codex (xhigh thinking)"
GPT_5_4_MINI_LOW = "gpt-5.4-mini (low thinking)"
GPT_5_4_2026_03_05_NONE = "gpt-5.4-2026-03-05 (no thinking)"
GPT_5_4_2026_03_05_LOW = "gpt-5.4-2026-03-05 (low thinking)"
@@ -23,8 +18,6 @@ class Llm(Enum):
GPT_5_5_XHIGH = "gpt-5.5 (xhigh thinking)"
# Claude
CLAUDE_SONNET_4_6 = "claude-sonnet-4-6"
- CLAUDE_4_5_SONNET_2025_09_29 = "claude-sonnet-4-5-20250929"
- CLAUDE_4_5_OPUS_2025_11_01 = "claude-opus-4-5-20251101"
CLAUDE_OPUS_4_6 = "claude-opus-4-6"
CLAUDE_OPUS_4_8_LOW = "claude-opus-4-8 (low effort)"
CLAUDE_OPUS_4_8_MEDIUM = "claude-opus-4-8 (medium effort)"
@@ -53,15 +46,10 @@ class Completion(TypedDict):
# models elsewhere in the codebase.
MODEL_PROVIDER: dict[Llm, str] = {
# OpenAI models
- Llm.GPT_4_1_2025_04_14: "openai",
Llm.GPT_5_2_CODEX_LOW: "openai",
Llm.GPT_5_2_CODEX_MEDIUM: "openai",
Llm.GPT_5_2_CODEX_HIGH: "openai",
Llm.GPT_5_2_CODEX_XHIGH: "openai",
- Llm.GPT_5_3_CODEX_LOW: "openai",
- Llm.GPT_5_3_CODEX_MEDIUM: "openai",
- Llm.GPT_5_3_CODEX_HIGH: "openai",
- Llm.GPT_5_3_CODEX_XHIGH: "openai",
Llm.GPT_5_4_MINI_LOW: "openai",
Llm.GPT_5_4_2026_03_05_NONE: "openai",
Llm.GPT_5_4_2026_03_05_LOW: "openai",
@@ -71,8 +59,6 @@ MODEL_PROVIDER: dict[Llm, str] = {
Llm.GPT_5_5_XHIGH: "openai",
# Anthropic models
Llm.CLAUDE_SONNET_4_6: "anthropic",
- Llm.CLAUDE_4_5_SONNET_2025_09_29: "anthropic",
- Llm.CLAUDE_4_5_OPUS_2025_11_01: "anthropic",
Llm.CLAUDE_OPUS_4_6: "anthropic",
Llm.CLAUDE_OPUS_4_8_LOW: "anthropic",
Llm.CLAUDE_OPUS_4_8_MEDIUM: "anthropic",
@@ -97,15 +83,10 @@ ANTHROPIC_MODELS = {m for m, p in MODEL_PROVIDER.items() if p == "anthropic"}
GEMINI_MODELS = {m for m, p in MODEL_PROVIDER.items() if p == "gemini"}
OPENAI_MODEL_CONFIG: dict[Llm, dict[str, str]] = {
- Llm.GPT_4_1_2025_04_14: {"api_name": "gpt-4.1-2025-04-14"},
Llm.GPT_5_2_CODEX_LOW: {"api_name": "gpt-5.2-codex", "reasoning_effort": "low"},
Llm.GPT_5_2_CODEX_MEDIUM: {"api_name": "gpt-5.2-codex", "reasoning_effort": "medium"},
Llm.GPT_5_2_CODEX_HIGH: {"api_name": "gpt-5.2-codex", "reasoning_effort": "high"},
Llm.GPT_5_2_CODEX_XHIGH: {"api_name": "gpt-5.2-codex", "reasoning_effort": "xhigh"},
- Llm.GPT_5_3_CODEX_LOW: {"api_name": "gpt-5.3-codex", "reasoning_effort": "low"},
- Llm.GPT_5_3_CODEX_MEDIUM: {"api_name": "gpt-5.3-codex", "reasoning_effort": "medium"},
- Llm.GPT_5_3_CODEX_HIGH: {"api_name": "gpt-5.3-codex", "reasoning_effort": "high"},
- Llm.GPT_5_3_CODEX_XHIGH: {"api_name": "gpt-5.3-codex", "reasoning_effort": "xhigh"},
Llm.GPT_5_4_MINI_LOW: {"api_name": "gpt-5.4-mini", "reasoning_effort": "low"},
Llm.GPT_5_4_2026_03_05_NONE: {
"api_name": "gpt-5.4-2026-03-05",
diff --git a/backend/run_evals.py b/backend/run_evals.py
index 8b5d7102..d637bea7 100644
--- a/backend/run_evals.py
+++ b/backend/run_evals.py
@@ -11,48 +11,5 @@ async def main():
await run_image_evals()
-# async def text_main():
-# OUTPUT_DIR = EVALS_DIR + "/outputs"
-
-# GENERAL_TEXT_V1 = [
-# "Login form",
-# "Simple notification",
-# "button",
-# "saas dashboard",
-# "landing page for barber shop",
-# ]
-
-# tasks: list[Coroutine[Any, Any, str]] = []
-# for prompt in GENERAL_TEXT_V1:
-# for n in range(N): # Generate N tasks for each input
-# if n == 0:
-# task = generate_code_for_text(
-# text=prompt,
-# stack=STACK,
-# model=Llm.CLAUDE_4_5_SONNET_2025_09_29,
-# )
-# else:
-# task = generate_code_for_text(
-# text=prompt, stack=STACK, model=Llm.GPT_4_1_2025_04_14
-# )
-# tasks.append(task)
-
-# print(f"Generating {len(tasks)} codes")
-
-# results = await asyncio.gather(*tasks)
-
-# os.makedirs(OUTPUT_DIR, exist_ok=True)
-
-# for i, content in enumerate(results):
-# # Calculate index for filename and output number
-# eval_index = i // N
-# output_number = i % N
-# filename = GENERAL_TEXT_V1[eval_index]
-# # File name is derived from the original filename in evals with an added output number
-# output_filename = f"{os.path.splitext(filename)[0]}_{output_number}.html"
-# output_filepath = os.path.join(OUTPUT_DIR, output_filename)
-# with open(output_filepath, "w") as file:
-# file.write(content)
-
if __name__ == "__main__":
asyncio.run(main())
diff --git a/backend/tests/test_openai_turn_input_logging.py b/backend/tests/test_openai_turn_input_logging.py
index da0cf6a6..d000eeba 100644
--- a/backend/tests/test_openai_turn_input_logging.py
+++ b/backend/tests/test_openai_turn_input_logging.py
@@ -57,7 +57,7 @@ def test_openai_turn_input_logger_preserves_full_large_payloads(
) -> None:
monkeypatch.setenv("LOGS_PATH", str(tmp_path))
- logger = OpenAITurnInputLogger(model=Llm.GPT_5_3_CODEX_LOW, enabled=True)
+ logger = OpenAITurnInputLogger(model=Llm.GPT_5_2_CODEX_LOW, enabled=True)
logger.record_turn_input(
[
{
diff --git a/backend/tests/test_token_usage.py b/backend/tests/test_token_usage.py
index 4cb4eaea..c9106f78 100644
--- a/backend/tests/test_token_usage.py
+++ b/backend/tests/test_token_usage.py
@@ -281,7 +281,6 @@ class TestAnthropicExtract:
class TestModelPricing:
def test_known_models_have_pricing(self) -> None:
for name in [
- "gpt-4.1-2025-04-14",
"gpt-5.2-codex",
"claude-opus-4-6",
"claude-sonnet-4-6",
@@ -303,5 +302,5 @@ class TestModelPricing:
assert MODEL_PRICING[name].cache_write > 0
def test_openai_gemini_no_cache_write(self) -> None:
- for name in ["gpt-4.1-2025-04-14", "gemini-3-flash-preview"]:
+ for name in ["gpt-5.2-codex", "gemini-3-flash-preview"]:
assert MODEL_PRICING[name].cache_write == 0.0
diff --git a/frontend/src/App.tsx b/frontend/src/App.tsx
index 1361a3f6..3f0307fc 100644
--- a/frontend/src/App.tsx
+++ b/frontend/src/App.tsx
@@ -93,7 +93,7 @@ function App() {
isImageGenerationEnabled: true,
editorTheme: EditorTheme.COBALT,
generatedCodeConfig: Stack.HTML_TAILWIND,
- codeGenerationModel: CodeGenerationModel.CLAUDE_4_5_OPUS_2025_11_01,
+ codeGenerationModel: CodeGenerationModel.CLAUDE_OPUS_4_6,
selectedDesignSystemId: null,
// Only relevant for hosted version
isTermOfServiceAccepted: false,
diff --git a/frontend/src/lib/models.ts b/frontend/src/lib/models.ts
index b429a1da..f5121363 100644
--- a/frontend/src/lib/models.ts
+++ b/frontend/src/lib/models.ts
@@ -8,17 +8,11 @@ export enum CodeGenerationModel {
CLAUDE_OPUS_4_8_MAX = "claude-opus-4-8 (max effort)",
CLAUDE_OPUS_4_6 = "claude-opus-4-6",
CLAUDE_SONNET_4_6 = "claude-sonnet-4-6",
- CLAUDE_4_5_OPUS_2025_11_01 = "claude-opus-4-5-20251101",
- CLAUDE_4_5_SONNET_2025_09_29 = "claude-sonnet-4-5-20250929",
GPT_5_5_XHIGH = "gpt-5.5 (xhigh thinking)",
GPT_5_2_CODEX_LOW = "gpt-5.2-codex (low thinking)",
GPT_5_2_CODEX_MEDIUM = "gpt-5.2-codex (medium thinking)",
GPT_5_2_CODEX_HIGH = "gpt-5.2-codex (high thinking)",
GPT_5_2_CODEX_XHIGH = "gpt-5.2-codex (xhigh thinking)",
- GPT_5_3_CODEX_LOW = "gpt-5.3-codex (low thinking)",
- GPT_5_3_CODEX_MEDIUM = "gpt-5.3-codex (medium thinking)",
- GPT_5_3_CODEX_HIGH = "gpt-5.3-codex (high thinking)",
- GPT_5_3_CODEX_XHIGH = "gpt-5.3-codex (xhigh thinking)",
GPT_5_4_MINI_LOW = "gpt-5.4-mini (low thinking)",
GEMINI_3_FLASH_PREVIEW_HIGH = "gemini-3-flash-preview (high thinking)",
GEMINI_3_FLASH_PREVIEW_MINIMAL = "gemini-3-flash-preview (minimal thinking)",
@@ -50,22 +44,9 @@ export const CODE_GENERATION_MODEL_DESCRIPTIONS: {
"gpt-5.2-codex (xhigh thinking)": {
name: "GPT 5.2 Codex (xhigh)",
},
- "gpt-5.3-codex (low thinking)": {
- name: "GPT 5.3 Codex (low)",
- },
- "gpt-5.3-codex (medium thinking)": {
- name: "GPT 5.3 Codex (medium)",
- },
- "gpt-5.3-codex (high thinking)": {
- name: "GPT 5.3 Codex (high)",
- },
- "gpt-5.3-codex (xhigh thinking)": {
- name: "GPT 5.3 Codex (xhigh)",
- },
"gpt-5.4-mini (low thinking)": {
name: "GPT 5.4 Mini (low)",
},
- "claude-opus-4-5-20251101": { name: "Claude Opus 4.5" },
"claude-opus-4-8 (low effort)": {
name: "Claude Opus 4.8 (low)",
},
@@ -83,7 +64,6 @@ export const CODE_GENERATION_MODEL_DESCRIPTIONS: {
},
"claude-opus-4-6": { name: "Claude Opus 4.6" },
"claude-sonnet-4-6": { name: "Claude Sonnet 4.6" },
- "claude-sonnet-4-5-20250929": { name: "Claude Sonnet 4.5" },
"gemini-3.5-flash (high thinking)": {
name: "Gemini 3.5 Flash (high)",
},
diff --git a/frontend/src/tests/qa.test.ts b/frontend/src/tests/qa.test.ts
index c2c31868..cbecee72 100644
--- a/frontend/src/tests/qa.test.ts
+++ b/frontend/src/tests/qa.test.ts
@@ -35,7 +35,7 @@ const stacks = process.env.QA_STACKS
? process.env.QA_STACKS.split(",").map((stack) => stack.trim() as Stack)
: defaultStacks;
-const defaultModels = [CodeGenerationModel.CLAUDE_4_5_OPUS_2025_11_01];
+const defaultModels = [CodeGenerationModel.CLAUDE_OPUS_4_6];
const models = process.env.QA_MODELS
? process.env.QA_MODELS.split(",").map((model) => model.trim())
: defaultModels;