coder-eval 0.8.5__tar.gz → 0.8.6__tar.gz

{coder_eval-0.8.5 → coder_eval-0.8.6}/CHANGELOG.md

@@ -2,6 +2,15 @@
 
 <!-- version list -->
 
+## v0.8.6 (2026-07-20)
+
+### Features
+
+- Add Sonnet 5 + GPT-5.6 pricing; default antigravity to gemini-3.5-flash
+  ([#31](https://github.com/UiPath/coder_eval/pull/31),
+  [`4c40acd`](https://github.com/UiPath/coder_eval/commit/4c40acd64482e9ffcec66b61ac2aab57b5c89f6f))
+
+
 ## v0.8.5 (2026-07-20)
 
 ### Bug Fixes

{coder_eval-0.8.5 → coder_eval-0.8.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: coder-eval
-Version: 0.8.5
+Version: 0.8.6
 Summary: Evaluate, benchmark, and A/B-test AI coding agents (Claude Code, Codex, Gemini/Antigravity) with sandboxed, reproducible YAML task suites.
 Project-URL: Homepage, https://github.com/UiPath/coder_eval
 Project-URL: Repository, https://github.com/UiPath/coder_eval

{coder_eval-0.8.5 → coder_eval-0.8.6}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "coder-eval"
-version = "0.8.5"
+version = "0.8.6"
 description = "Evaluate, benchmark, and A/B-test AI coding agents (Claude Code, Codex, Gemini/Antigravity) with sandboxed, reproducible YAML task suites."
 readme = "README.md"
 license = "Apache-2.0"

{coder_eval-0.8.5 → coder_eval-0.8.6}/src/coder_eval/__init__.py

@@ -1,3 +1,3 @@
 """coder_eval - A framework for evaluating AI coding agents."""
 
-__version__ = "0.8.5"
+__version__ = "0.8.6"

{coder_eval-0.8.5 → coder_eval-0.8.6}/src/coder_eval/agents/antigravity_agent.py

@@ -93,10 +93,10 @@ def _harness_spawn_lock() -> asyncio.Lock:
 
 
 # Recommended Gemini coding model when a task pins no ``agent.model`` and neither
-# ``--model`` nor ``ANTIGRAVITY_MODEL`` is set. Gemini 3.1 Pro is the current
-# flagship agentic-coding model (the earlier ``gemini-3-pro-preview`` is retired);
-# ``medium`` thinking is its recommended daily-driver default.
-_DEFAULT_MODEL = "gemini-3.1-pro-preview"
+# ``--model`` nor ``ANTIGRAVITY_MODEL`` is set. Gemini 3.5 Flash is Antigravity 2.0's
+# default coding model (2026-05) — it outperforms the older Gemini 3.1 Pro on coding /
+# agentic benchmarks while running faster; ``medium`` thinking is its daily-driver default.
+_DEFAULT_MODEL = "gemini-3.5-flash"
 
 # Antigravity builtin tool name -> canonical Claude-ish tool name, so cross-agent
 # success criteria (command_executed / commands_efficiency / skill_triggered) and

{coder_eval-0.8.5 → coder_eval-0.8.6}/src/coder_eval/pricing.py

@@ -28,6 +28,8 @@ _PRICING: dict[str, ModelPricing] = {
     "claude-opus-4-6-20250514": ModelPricing(15.0, 75.0, 18.75, 1.50),
     "claude-opus-4-5-20251101": ModelPricing(15.0, 75.0, 18.75, 1.50),
     "claude-opus-4-20250514": ModelPricing(15.0, 75.0, 18.75, 1.50),
+    # Claude 5 Sonnet (2026-06-30 on Bedrock): standard $3/$15 (promo $2/$10 thru 2026-08-31).
+    "claude-sonnet-5": ModelPricing(3.0, 15.0, 3.75, 0.30),
     # Claude 4.6 / 4.5 / 4 Sonnet
     "claude-sonnet-4-6": ModelPricing(3.0, 15.0, 3.75, 0.30),
     "claude-sonnet-4-6-20250514": ModelPricing(3.0, 15.0, 3.75, 0.30),
@@ -62,6 +64,18 @@ _PRICING: dict[str, ModelPricing] = {
     # for very-large-context runs reads low. Fine for typical eval tasks; revisit
     # if benchmarking large-context Codex runs.
     "gpt-5.5": ModelPricing(5.0, 30.0, 5.0, 0.50),
+    # gpt-5.5-pro / gpt-5.4-pro: $30/M in, $180/M out; pro tiers offer NO prompt
+    # caching (cache_read nominal, never billed since pro doesn't cache).
+    "gpt-5.5-pro": ModelPricing(30.0, 180.0, 30.0, 3.0),
+    "gpt-5.4-pro": ModelPricing(30.0, 180.0, 30.0, 3.0),
+    # gpt-5.4 economy tiers: mini $0.75/$4.50, nano $0.20/$1.25.
+    "gpt-5.4-mini": ModelPricing(0.75, 4.5, 0.75, 0.075),
+    "gpt-5.4-nano": ModelPricing(0.20, 1.25, 0.20, 0.02),
+    # GPT-5.6 family (2026-07-09): sol flagship / terra balanced (Codex default) / luna
+    # economy. Terra matches gpt-5.4's rate; sol matches gpt-5.5.
+    "gpt-5.6-sol": ModelPricing(5.0, 30.0, 6.25, 0.50),
+    "gpt-5.6-terra": ModelPricing(2.5, 15.0, 3.125, 0.25),
+    "gpt-5.6-luna": ModelPricing(1.0, 6.0, 1.25, 0.10),
     # Google Gemini (AntigravityAgent, via the Gemini Developer API). Per-MTok
     # rates from ai.google.dev/gemini-api/docs/pricing (2026). Gemini bills no
     # separate cache-WRITE fee, so cache_write == input (the agent maps

{coder_eval-0.8.5 → coder_eval-0.8.6}/uv.lock

@@ -449,7 +449,7 @@ wheels = [
 
 [[package]]
 name = "coder-eval"
-version = "0.8.5"
+version = "0.8.6"
 source = { editable = "." }
 dependencies = [
     { name = "anthropic" },