Grok 4.3 launches: xAI's new flagship, 37.5% input / 58.3% output price cuts

xAI shipped Grok 4.3 on April 30, 2026: Intelligence Index 53, GDPval-AA ELO 1500, τ²-Bench 98%, IFBench 81%, 159 tokens/s. Listed at just $1.25 / $2.50 per 1M tokens, available on all common groups, ~85% of upstream pricing after the 10% recharge bonus.

Key Highlights

• xAI’s new flagship — Grok 4.3 (April 30, 2026), positioned for “lowest hallucination rate, strongest agentic tool calling, best instruction following”
• Major price cuts — input down 37.5% and output down 58.3% vs. Grok 4.20 0309 v2; full Intelligence Index eval costs about 20% less
• Blazing fast — 159 tokens/s output, well above the 64.6 t/s median for same-tier reasoning models
• Million-token context + multimodal — 1M context window, text + image input, always-on reasoning (cannot be disabled)
• All common groups available — Default, SVIP, etc. all open; deposit $100 get 10% bonus, ~85% of xAI’s upstream pricing

Background

On April 30, 2026 (UTC+8), xAI shipped Grok 4.3 as the next flagship after Grok 4.20. The headline isn’t a benchmark sweep — it’s a cost-performance leap: a score of 53 on the Artificial Analysis Intelligence Index with the full eval suite costing only $395 to run, ~20% cheaper than Grok 4.20 0309 v2.

xAI’s official positioning calls out “industry-leading non-hallucination rate, agentic tool calling, and instruction following” — three dimensions all targeted at production-grade agent workloads. On the Intelligence Index, Grok 4.3 edges past Claude Sonnet 4.6; the single biggest jump is on GDPval-AA, where ELO leapt from Grok 4.20 v2’s 1179 straight to 1500 — a 321-point one-generation gain.

APIYI has shipped Grok 4.3 with all common user groups (Default, SVIP, etc.) open — pricing is friendly, blast radius is contained, no need to gate it like GPT-5.5 Pro. Stack the 10% deposit bonus and effective cost lands around 85% of xAI’s upstream rate.

Detailed Breakdown

Core Capabilities

Industry-leading agent tool calling

τ²-Bench Telecom 98%, IFBench 81%; stable on complex tool chains and multi-step agents

Lowest-class hallucination rate

AA-Omniscience Accuracy +8 vs. previous; xAI claims industry-leading non-hallucination rate

Very fast output

159 tokens/s — 2.5× the 64.6 t/s median for same-tier reasoning models

Million-token context + multimodal

1M tokens context, text + image input, always-on reasoning

Benchmarks

Benchmark	Grok 4.3	Grok 4.20 v2	Δ
Intelligence Index	53	49	+4
GDPval-AA (ELO)	1500	1179	+321
τ²-Bench Telecom	98%	—	Industry-leading
IFBench	81%	—	Instruction following
AA-Omniscience Accuracy	+8 vs. v2	baseline	Accuracy lift
Output speed	159 t/s	lower	Big jump
Cost to run Index	$395	~$494	-20%

Sources: xAI official announcement, Artificial Analysis independent benchmarks (April 30, 2026). The Intelligence Index aggregates multiple benchmarks and is a useful proxy for overall model intelligence.

Tech Specs

Parameter	Grok 4.3
Model name	grok-4.3
Context window	1,000,000 tokens
Max output	No explicit cap (within remaining context)
Input modalities	Text + image
Reasoning mode	Always-on reasoning (cannot be disabled)
Output speed	~159 tokens/s
API endpoint	/v1/chat/completions
Available groups	Default, SVIP, and other common groups

Practical Use

Recommended Scenarios

Grok 4.3’s combo — high speed + low price + strong agent — fits well for:

1. Production agent workflows — high-frequency tool calls, multi-step planning, scenarios that demand stable instruction following
2. Large-scale document processing — 1M context + 159 t/s makes long-doc summarization and cross-file audits run fast
3. Customer support / telecom dialogues — 98% on τ²-Bench Telecom suggests strong performance on ticketing and tech-support tool-calling tasks
4. Multimodal analysis — image + text mixed input for screenshot analysis, chart reading
5. Cost-sensitive high-QPS apps — low unit price + fast speed makes Grok 4.3 a reasonable substitute for GPT-5.4 / Claude Sonnet in some flows

Code Examples

Standard call

from openai import OpenAI

client = OpenAI(
    api_key="your-apiyi-key",
    base_url="https://api.apiyi.com/v1"
)

response = client.chat.completions.create(
    model="grok-4.3",
    messages=[
        {"role": "user", "content": "Analyze this customer ticket and propose a resolution"}
    ],
    max_tokens=4096
)

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

Multimodal call (image + text)

response = client.chat.completions.create(
    model="grok-4.3",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What's the cause of this error in the screenshot?"},
                {"type": "image_url", "image_url": {"url": "https://example.com/error.png"}}
            ]
        }
    ]
)

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

Agent tool calling

tools = [
    {
        "type": "function",
        "function": {
            "name": "query_order_status",
            "description": "Look up a customer order status",
            "parameters": {
                "type": "object",
                "properties": {"order_id": {"type": "string"}},
                "required": ["order_id"]
            }
        }
    }
]

response = client.chat.completions.create(
    model="grok-4.3",
    messages=[{"role": "user", "content": "Check the status of order #A1024"}],
    tools=tools,
    tool_choice="auto"
)

Best Practices

1. Pick it for agents — τ²-Bench / IFBench data points to a clear edge in tool calling and instruction following over generic models
2. Mind the 200K context tier — past 200K, input/output prices double; control prompt length on long-doc workflows
3. Always-on reasoning is default — reasoning tokens count against your output bill; even simple tasks incur some reasoning spend
4. Speed-first chat — 159 t/s throughput makes Grok 4.3 great for “stream-and-render” real-time chat UX
5. Stack the recharge bonus — APIYI’s deposit $100 / +10% offer effectively brings cost to ~85% of upstream

Pricing & Availability

Pricing (Tiered)

Context range	Input	Output	Notes
0 – 200K tokens	$1.25 / 1M tokens	$2.50 / 1M tokens	Standard tier, matches xAI upstream
200K – ∞ tokens	$2.50 / 1M tokens	$5.00 / 1M tokens	Long-context tier, 2x premium

Price comparison with recent models

Model	Input	Output	Speed	Position
Grok 4.3	$1.25	$2.50	159 t/s	High-speed agent value pick
Grok 4.20	$2.00	$6.00	lower	Previous flagship
Claude Sonnet 4.6	$3.00	$15.00	medium	Coding generalist
GPT-5.5	$5.00	$30.00	medium	Frontier reasoning
Gemini 3 Pro	$2.00	$12.00	higher	Multimodal

At $1.25 / $2.50, Grok 4.3 has essentially no same-tier rival — comparable-performance models cost 2-10× more. Cost-performance is the headline.

Stack with site recharge promotions

Deposit $100 get 10% bonus, ~85% of upstream

APIYI offers a deposit $100 / +10% bonus. Stacked on listed pricing, effective cost lands around 85% of xAI’s upstream rate — better the longer you use it.

Available models & groups

Model name	Channel	Available groups	Notes
grok-4.3	xAI official relay	Default, SVIP, and other common groups	Latest, auto-tracks upstream

Access

• Site: apiyi.com
• API endpoint: https://api.apiyi.com/v1
• OpenAI SDK compatible — just swap base_url, api_key, and set model to grok-4.3

Summary & Recommendations

Grok 4.3 is xAI’s most cost-effective flagship to date: $1.25 / $2.50 + 159 t/s + 1M context + strong agent skills — a combination that has essentially no competitor in the $1-3 input-price band.

Worth trying / migrating to Grok 4.3 when:

• You’re already running Grok 4.20 / Claude Sonnet 4.6 / GPT-5.4 mini for agent tasks — evaluate migration first
• High-frequency tool calling, customer-support automation, telecom-style ticketing
• Cost-sensitive, high-QPS real-time chat products
• Workloads needing image input multimodal analysis

Don’t switch when:

• You’re already on GPT-5.5 / Claude Opus 4.7 for the hardest reasoning (Grok 4.3’s intelligence ceiling is lower)
• You depend heavily on OpenAI/Anthropic-specific API behaviors (e.g., OpenAI’s strict function schema)
• You’re stable on GPT-5.4 / Gemini 3 Pro and the migration upside is marginal

APIYI has Grok 4.3 open on Default, SVIP, and other common groups — friendly pricing, fast speed, contained risk. After the 10% recharge bonus, effective cost is roughly 85% of xAI’s upstream. Recommend a small-scale A/B on your real workloads before rolling out broadly.

Sources: xAI official API docs (docs.x.ai), Artificial Analysis independent benchmarks, VentureBeat and The Decoder coverage. Data captured: May 3, 2026 (UTC+8).