fix(pricing): add missing Kiro model pricing rows (Sonnet 4.6, auto-kiro, deepseek-3.2, minimax-m2.5, glm-5)

[artickc]

Summary

The kiro table in DEFAULT_PRICING (src/shared/constants/pricing.ts) was missing several models that the Kiro provider registry actually serves. Because getPricingForModel() does an exact-key lookup, those models resolved to null and their usage cost was reported as $0.00.

Most visibly, claude-sonnet-4.6 had no pricing row (the table only had claude-sonnet-4.5 and claude-sonnet-4).

What was missing

Diffing getModelsByProviderId("kiro") against DEFAULT_PRICING.kiro surfaced 5 served-but-unpriced model ids:

Model id (registry)	Note
claude-sonnet-4.6	no row at all
auto-kiro	table only had auto
deepseek-3.2	table only had deepseek-v3.2
minimax-m2.5	table only had minimax-m2.1
glm-5	no row under kiro

Changes

• Add the 5 missing rows to DEFAULT_PRICING.kiro:
  • claude-sonnet-4.6 mirrors claude-sonnet-4.5 ($3 in / $15 out)
  • deepseek-3.2 and auto-kiro mirror their existing siblings
  • minimax-m2.5 = $0.27 / $0.95 (matches the MiniMax table)
  • glm-5 = $1.0 / $3.2 (matches the GLM table)
• Add a regression test in tests/unit/catalog-updates-v3x.test.ts asserting that every model returned by getModelsByProviderId("kiro") resolves a numeric pricing row, so this class of $0.00 bug is caught going forward.

Testing

node --import tsx --import ./open-sse/utils/setupPolyfill.ts --test tests/unit/catalog-updates-v3x.test.ts
# 6 pass, 0 fail

[gemini-code-assist]

Code Review

This pull request adds pricing configurations for several new models, including claude-sonnet-4.6, deepseek-3.2, minimax-m2.5, glm-5, and auto-kiro, and introduces unit tests to validate these updates. The reviewer identified that the reasoning cost for claude-sonnet-4.6 is incorrectly set to 15.0 instead of 22.5, and recommended updating both the pricing constant and its corresponding regression test to reflect the correct value.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

The reasoning cost for claude-sonnet-4.6 is set to 15.0 (matching Sonnet 4.5), but Sonnet 4.6 has a 1.5x reasoning multiplier ($22.50 per million tokens) as defined in CLAUDE_SONNET_46_PRICING and used by other providers (e.g., anthropic and ag). Let's update this to 22.5 to ensure accurate cost tracking.

Suggested change

	"claude-sonnet-4.6": {
	input: 3.0,
	output: 15.0,
	cached: 1.5,
	reasoning: 15.0,
	cache_creation: 3.0,
	},
	"claude-sonnet-4.6": {
	input: 3.0,
	output: 15.0,
	cached: 1.5,
	reasoning: 22.5,
	cache_creation: 3.0,
	},

[gemini-code-assist]

Let's update the regression test to also assert the corrected reasoning cost of 22.5 for claude-sonnet-4.6.

  const sonnet46 = getPricingForModel("kiro", "claude-sonnet-4.6") as {
    input: number;
    output: number;
    reasoning?: number;
  } | null;
  assert.ok(sonnet46, "kiro pricing must include claude-sonnet-4.6");
  assert.equal(sonnet46?.input, 3.0);
  assert.equal(sonnet46?.output, 15.0);
  assert.equal(sonnet46?.reasoning, 22.5);

[diegosouzapw]

Merged into release/v3.8.25 🎉 — thanks @artickc! Spot-on catch: the missing kiro rows (esp. claude-sonnet-4.6) meant getPricingForModel() returned null and usage cost showed $0.00. Ships in v3.8.25. 🙌