fix: update xAI model metadata for grok-4.20-beta-0309-reasoning +2 more

[github-actions]

fix: update xAI model metadata for grok-4.20-beta-0309-reasoning +2 more

Closes #695

Source issue: #695

Summary

Field	Value
Provider	xai
Primary model	grok-4.20-beta-0309-reasoning
Changed models	grok-4.20-beta-0309-reasoning grok-4.20-beta-0309-non-reasoning grok-4.20-multi-agent-beta-0309
Added models	None
Updated models	grok-4.20-beta-0309-reasoning grok-4.20-beta-0309-non-reasoning grok-4.20-multi-agent-beta-0309
Verification sources	1 2 3

Verified metadata

Model	Display name	Parent	Providers	Format	Flavor	Token limits	Pricing	Lifecycle
grok-4.20-beta-0309-reasoning			xAI	openai	chat	input=2000000, output=2000000	in/out=1.25/2.5 per 1M; cache read=0.2 per 1M	multimodal=true; reasoning=true
grok-4.20-beta-0309-non-reasoning			xAI	openai	chat	input=2000000, output=2000000	in/out=1.25/2.5 per 1M; cache read=0.2 per 1M	multimodal=true
grok-4.20-multi-agent-beta-0309			xAI	openai	chat	input=2000000, output=2000000	in/out=1.25/2.5 per 1M; cache read=0.2 per 1M	multimodal=true; reasoning=true

Verification notes

Verification

• Source 1: xAI models page — lists all grok-4.20 chat models at $1.25/$2.50 per million tokens (input/output). Cached input at $0.20/M tokens. Context window: 1,000,000 tokens.
• Source 2: xAI grok-4.20-multi-agent-0309 detail page — confirms $1.25/$2.50 pricing, lists grok-4.20-multi-agent-beta-0309 as an alias.
• Source 3: xAI grok-4.20-beta-0309-reasoning detail page — confirms $1.25/$2.50 pricing and $0.20 cached input.
• Source 4: Internal consistency — the catalog's own non-beta entries already have $1.25/$2.50 pricing.
• sync_models (LiteLLM): No grok-4.20 entries found in the LiteLLM model_prices_and_context_window_backup.json reference catalog, so no cross-check values are available.
• Fields verified: input_cost_per_mil_tokens and output_cost_per_mil_tokens for all three models, confirmed by xAI official docs.
• Fields unchanged: format, flavor, multimodal, input_cache_read_cost_per_mil_tokens, reasoning, max_input_tokens, max_output_tokens, available_providers — these are not changed by this update and are kept at their existing values.
• Note on token limits: The xAI docs list a 1,000,000 token context window for all grok-4.20 models, but the existing catalog entries have max_input_tokens: 2000000. This discrepancy is pre-existing and out of scope for this pricing-only update.

sync_models vs proposed update

sync_models cross-check found differences. Official provider verification was used for the applied values, and sync_models discrepancies are listed below for review.

Model	Field	Proposed update	sync_models	sync_models source models
grok-4.20-beta-0309-reasoning	input_cost_per_mil_tokens	1.25	2	xai/grok-4.20-beta-0309-reasoning
grok-4.20-beta-0309-reasoning	output_cost_per_mil_tokens	2.5	6	xai/grok-4.20-beta-0309-reasoning
grok-4.20-beta-0309-non-reasoning	input_cost_per_mil_tokens	1.25	2	xai/grok-4.20-beta-0309-non-reasoning
grok-4.20-beta-0309-non-reasoning	output_cost_per_mil_tokens	2.5	6	xai/grok-4.20-beta-0309-non-reasoning
grok-4.20-multi-agent-beta-0309	input_cost_per_mil_tokens	1.25	2	xai/grok-4.20-multi-agent-beta-0309
grok-4.20-multi-agent-beta-0309	output_cost_per_mil_tokens	2.5	6	xai/grok-4.20-multi-agent-beta-0309

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
ai-proxy	Ready	Preview, Comment	May 31, 2026 2:41pm