xAI Grok 4.3 on Amazon Bedrock (June 2026): Mantle Routing and Rate-Card Math

AWS lifecycle notice (June 30, 2026) — Amazon Bedrock Agents Classic is in maintenance for new customers after July 30, 2026. Net-new agent builds should use Bedrock AgentCore. Full matrix: lifecycle roundup.

On June 15, 2026, AWS announced xAI Grok 4.3 general availability on Amazon Bedrock, served through the Bedrock Mantle endpoint. Published on-demand list rates: $1.25 / $2.50 per 1M input/output tokens, with $0.20 / 1M cached input (re-verify on Bedrock pricing before you lock a budget).

This is a third frontier option on Mantle beside OpenAI GPT-5.6 and Anthropic Claude — not a drop-in for every Converse-based agent you already run. If you standardize on Bedrock for procurement and IAM, Grok 4.3 is worth a routing slot when long context and aggressive input pricing matter more than staying on a Converse-native path.

Successor: Grok 4.7 (September 28, 2026)

What AWS changed: on September 28, 2026, Grok 4.7 became available on Bedrock. Per the model card: text and image input, 500K context, reasoning effort low / medium / high / xhigh, and US Geo (us.xai.grok-4.7) or Global (global.xai.grok-4.7) cross-Region inference on bedrock-runtime — In-Region is not supported. Responses, Chat Completions, and Converse all work. Standard rates match 4.6: $2.20 / $6.60 / $0.55 cache (US Geo) and $2.00 / $6.00 / $0.50 (Global) per 1M tokens. Priority is billed at 1.75x and Flex at 0.5x.

What the launch claims: better handling of mixed documents, more reliable repo-scale coding with planning and error recovery, and stronger browser-use agents for form fills and portal navigation. AWS cites Artificial Analysis moving the Intelligence Index from 44 to 46 and the Coding Agent Index from 47 to 56 versus 4.6.

The catch: the same source shows average output per task roughly doubling, from about 38K to 81K tokens. Same rate card, more tokens per task — a lane that was cheap on 4.6 can cost more on 4.7. Cap maxTokens, start reasoning effort at low or medium, and measure cost per completed task in a frozen-prompt bakeoff before moving a 4.6 lane.

Previous successor: Grok 4.6 (August 19, 2026)

Previous approach: Grok 4.3 on Mantle only — Chat Completions / Responses, three US Regions, no CRIS profile IDs, no Converse.

What AWS changed: on August 19, 2026, Grok 4.6 reached GA on Bedrock with US Geo (us.xai.grok-4.6) and Global (global.xai.grok-4.6) cross-Region inference on bedrock-runtime, plus Converse, Responses, and Chat Completions. Mantle still accepts xai.grok-4.6 without CRIS. Official model card: 500K context; Standard $2.20 / $6.60 / $0.55 cache (In-Region and US Geo) and $2.00 / $6.00 / $0.50 (Global) per 1M tokens.

What this enables: xAI in a Converse-native agent or Guardrails composition path, and data-residency-aware CRIS without standing up a Mantle-only fork.

When it matters: you already want Grok behavior and Converse or geo/global throughput.

When it does not: a 4.3 Mantle lane that already meets quality at $1.25 / $2.50. 4.6 is a capability upgrade, not a price cut. Do not treat 4.6 as a 1M-context successor to 4.3’s ~1M window.

This post keeps the 4.3 rate-card math below. Recalculate 4.6 on the model card before you rewrite a budget.

What changed on June 15, 2026

Per AWS at GA:

Implication: teams on bedrock-runtime Converse (Agents Classic, many LangChain paths, Bedrock Flows) cannot swap model IDs silently. Mantle is a separate invoke surface — same billing account, different client configuration and IAM actions.

First-party pricing math (not a client silhouette)

No anonymized engagement is cited for Grok 4.3 adoption. The numbers below are reproducible arithmetic on the published Bedrock rate card — one illustrative monthly shape, labeled as math.

Illustrative month: 10M input tokens + 1M output tokens (agentic production mix with moderate output).

Direction: Grok 4.3 is −$30.00 / month (−67%) vs Sonnet 5 and −$20.20 / month (−57%) vs Terra on identical token counts. Luna is still ~4× cheaper on this mix — use Luna for high-QPS classification, not as a Grok substitute for 1M-context reasoning lanes.

Download the full comparison: model-routing-cost-worksheet.csv.

Reproduce this — Open the worksheet CSV and replace the sample 10M/1M columns with your Cost Explorer / CUR input and output totals. Work the Mantle gotchas checklist before promoting a lane. Pair with Bedrock cost optimization: token budgets and model selection.

Opinionated routing: where Grok 4.3 belongs

We recommend: trial Grok 4.3 on new Mantle-native lanes that need long context (large codebases, multi-document RAG, extended chat history) and where list input cost dominates — after a frozen-prompt bakeoff against Terra and Sonnet 5. Keep Converse-native production on Sonnet 5 until Mantle migration is explicit. Keep high-volume thin tasks on Luna (see GPT-5.6 Luna & Terra pricing). Keep hardest agentic / governance-heavy work on Claude Opus 5 where ZDR-default and Guardrails composition matter (Opus 5 on Bedrock).

Trade-off you accept: lower $/MTok and 1M context vs Mantle-only invoke, no Converse tool path, and xAI-specific behavior you must re-validate on your prompt pack.

For OpenAI-on-Bedrock Mantle patterns (Responses API, Codex context), see OpenAI models + Codex on Bedrock.

What broke (pattern): Converse assumptions on a Mantle-only model

What broke — A team pointed an existing Bedrock Converse wrapper at xai.grok-4.3 and got ValidationException: model does not support Converse. Root cause: Grok 4.3 is Mantle-only at GA. Detection: CloudWatch ModelInvocationError on bedrock-runtime with zero Mantle traffic. Fix: migrate invoke to bedrock-mantle.REGION.api.aws/openai/v1, update IAM from runtime-only to Mantle actions, and re-test tool schemas — Chat Completions message format differs from Converse content blocks.

Secondary gotcha: context-tier pricing. List rates ($1.25 / $2.50) may apply only below a context threshold; input cost can roughly double above ~200K tokens on some Bedrock model cards — verify the xAI row before budgeting a 1M-context lane. See mantle-gotchas.md.

Invoke Grok 4.3 on Mantle

Context: OpenAI Python SDK against Bedrock Mantle, region us-east-1, model ID from the Grok 4.3 model card. Auth via Bedrock API key or SigV4 per your path.

# openai>=1.x; OPENAI_BASE_URL=https://bedrock-mantle.us-east-1.api.aws/openai/v1
from openai import OpenAI

client = OpenAI()  # OPENAI_API_KEY + OPENAI_BASE_URL from env

response = client.chat.completions.create(
    model='xai.grok-4.3',
    messages=[
        {'role': 'system', 'content': 'You are a concise support assistant.'},
        {'role': 'user', 'content': 'Classify: billing dispute on invoice #4421.'},
    ],
    # reasoning_effort='medium',  # none | low | medium | high — verify model card
)

print(response.choices[0].message.content)

Regions today: us-east-1, us-east-2, us-west-2. No geo/global inference IDs — capacity-plan inside those Regions. Converse on bedrock-runtime is not an fallback.

What to Do This Week

1. Confirm Grok 4.3 is enabled in your Bedrock model access console for the target Region. If the lane needs Converse or CRIS, evaluate Grok 4.6 (us.xai.grok-4.6 / global.xai.grok-4.6) as a separate bakeoff — list rates are higher.
2. Pull last 30 days of Bedrock token usage; split Converse vs Mantle usage types in CUR.
3. Recompute monthly $ with the worksheet CSV at Grok, Terra, Sonnet 5, and Luna rows.
4. Pick one long-context lane that is Mantle-ready; freeze 50–100 production prompts; score Grok vs incumbent on pass rate, p95 latency, and $ / completed task at reasoning_effort=none and medium.
5. Check context-tier pricing if any prompt exceeds ~200K input tokens.
6. Update IAM, secrets, and observability dashboards to tag Mantle model IDs — Converse-only Cost Explorer filters will miss Grok spend.
7. If you are net-new on agents post–July 30, 2026, route orchestration to AgentCore rather than Agents Classic.

What This Post Doesn’t Cover

• Grok 4.6 quality vs 4.3 on our internal prompt packs (no harness linked yet). Use the Grok 4.6 model card for live rates and API support.
• Batch, Provisioned Throughput, or Priority tier pricing for Grok 4.3 — confirm on the model card; GA notes reference on-demand only.
• xAI first-party API vs Bedrock residency/compliance comparison beyond pointing at AWS procurement consolidation.
• EU / non-US regional availability beyond the three US Regions AWS listed at GA.
• Guardrails + Knowledge Bases composition on the same Converse call path without a proxy layer.

Use the rate card + your CUR. Promote lane-by-lane.

Related reading

• OpenAI GPT-5.6 Luna & Terra Bedrock Price Cuts (July 2026)
• Claude Opus 5 on AWS Bedrock (July 2026)
• Claude Sonnet 5 on AWS (June 2026)
• AWS Bedrock Cost Optimization: Token Budgets and Model Selection
• Amazon Bedrock + OpenAI Models, Codex & Managed Agents
• Generative AI on AWS · Amazon Bedrock services

Frequently asked questions