[fireworks-ai] Fix Kimi K3 10x pricing error and update stale model prices

[akhilmmenon]

Summary

Fixes incorrect and stale pricing for Fireworks AI models in pricing/fireworks-ai.json.

• Kimi K3 — prices were 10x too high due to a unit conversion error (dollars-per-million entered directly instead of converting to cents-per-token)
• DeepSeek R1 — stale pricing ($8.00/$8.00 → $0.55/$2.19 per 1M)
• DeepSeek V4 Pro — updated to current 0813 version pricing (Standard & Priority)
• DeepSeek V4 Flash — updated to current 0731 version pricing
• GLM 5.2 (glm-5p2) — corrected stale cached input price ($0.26 → $0.14 per 1M)

Pricing References

Model	Source	Price ($/1M: input / cached / output)
Kimi K3	Fireworks Serverless Pricing	$3.00 / $0.30 / $15.00
DeepSeek R1	Firewor, LLM Cost Hub	$0.55 / — / $2.19
DeepSeek V4 Pro (0813)	Fireworks Serverless Pricing	Standard: $1.32 / $0.044 / $3.96, Priority: $1.65 / $0.055 / $4.95
DeepSeek V4 Flash (0731)	Fireworks Serverless Pricing	$0.22 / $0.007 / $0.66
GLM 5.2	Fireworks Serverless Pricing	$1.40 / $0.14 / $4.40

Pylon Ticket

https://app.usepylon.com/support/issues/views/4305f6d4-1eeb-4039-92aa-1867128acd4f?view=fs&issueNumber=8480

[Copilot]

🟡 Changes recommended

A separately exposed DeepSeek V4 Flash entry still uses outdated pricing.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

Corrects stale and misconverted Fireworks AI model pricing.

Changes:

• Fixes Kimi K3’s 10× pricing error.
• Updates DeepSeek R1, V4 Pro/Flash, and GLM 5.2 rates.

File summaries

File	Review
pricing/fireworks-ai.json	Critical: Update the explicit deepseek-v4-flash-0731 entry to the corrected input, cache, and output prices.

Review details

• Files reviewed: 1/1 changed files
• Comments generated: 1
• Review effort level: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.