Grok 4.20 review, pricing and verdict

Grok 4.20 is the pick for teams processing documents beyond 200K tokens who don't need audio input or on-premise deployment: legal, research, and coding teams that outgrew Grok 4.1's smaller context window. Its SWE-bench Verified score of 75% keeps it competitive with GPT-5.4 on agentic coding, not just document length.

A 2 million token context window and a 4-agent Mixture-of-Experts architecture define Grok 4.20, the xAI model that entered beta in February 2026. Four parallel agents debate and synthesize each response, extending well past the single-model context limits of earlier Grok releases.

Provider: xAI · Family: Grok 4

More about xAI on HokAI

Context window: 2,000,000 tokens

Input modalities: text, image, pdf, tool-calls · Output: text, tool-calls

About Grok 4.20

Grok 4.20 is xAI's fourth-generation flagship language model, released in public beta on February 17, 2026, with full general availability reaching the API on March 18, 2026. Built on a Mixture-of-Experts transformer backbone with an estimated 1.7 to 3 trillion total parameters, it is the most structurally distinct Grok release since the original series. Rather than routing queries to a single model instance, Grok 4.20 deploys a four-agent council at inference time: Grok (coordinator), Harper (research), Benjamin (math and code), and Lucas (synthesis and creativity). These agents run in parallel on shared weights, debate intermediate results through peer-review rounds, and synthesize a final answer. This design reduces hallucination rates from approximately 12% to 4.2% versus a single-model baseline, a 65% improvement documented in xAI internal testing.

On benchmark performance, Grok 4.20 scores 88.9% on GPQA Diamond for graduate-level scientific reasoning, placing it among the top tier of proprietary models, though GPT-5.4 (92.8%) and Claude Opus 4.6 (91.3%) hold the advantage on this specific axis. AIME 2025 math performance reaches approximately 95% for the standard four-agent variant; the 16-agent Heavy mode achieves a reported perfect score. On MMLU-Pro, the model scores 86.6%, consistent with the frontier cluster. SWE-bench Verified software engineering performance sits at 75%, essentially tied with GPT-5.4 at 74.9% and marginally ahead of Claude Opus 4.6 at 74%. Gemini 3.1 Pro trails at 68.3%, a meaningful gap for production coding workflows. LM Arena crowdsourced Elo ranged between 1505 and 1535 at provisional March 2026 measurement. IFBench instruction-following accuracy is 82.9%.

Grok 4.20 ships in three variants that differ primarily in context window and reasoning behavior. The multi-agent variant (grok-4.20-multi-agent-0309) supports a 2 million token context window, large enough to process entire multi-year codebases, full legal case archives, or extensive clinical trial datasets in a single API call. The reasoning and non-reasoning single-model variants (grok-4.20-0309-reasoning and grok-4.20-0309-non-reasoning) support a 1 million token context window. Output tokens count toward the declared context window rather than a separate cap. Community evaluations report strong long-context recall above 500K tokens, a marked improvement over Grok 4.1 which topped out at 131K and substantially ahead of GPT-5 at 128K.

On modalities and capabilities, Grok 4.20 accepts text and image inputs natively, with PDF handling available through the X platform file processing pipeline. Outputs are text and tool-call responses. Native function calling and JSON-structured output are supported on the xAI API with full parameter-level schema definitions. Real-time data from the X platform is integrated via a live search hook, extending the effective knowledge boundary beyond the static November 2024 training cutoff for current-events queries. The Heavy variant scales to 16 parallel agents for the most demanding decomposable tasks. Native audio input and output are absent from the 4.20 release; xAI has indicated these capabilities are on the roadmap for Grok 5.

As of May 2026, xAI cut Grok 4.20's per-token pricing well below its February 2026 launch rate, and cached input runs cheaper still than standard input (see the pricing FAQ for exact figures). To put current rates in context: processing a 500K-token research document costs approximately $0.63; a daily coding agent loop processing 1 million input tokens and 200K output tokens costs $1.75; a 1,000-turn customer-support session at 2K input and 500 output tokens per turn costs $3.75. Grok 4.20 undercuts both Claude Opus 4.8 and GPT-5 on input and output pricing. Running the Heavy variant's larger agent count multiplies total token consumption several times over versus standard mode, so high-volume batch pipelines are usually better served by the single-model reasoning variant.

Grok 4.20 is available via the xAI API at api.x.ai with API key authentication. Third-party gateway access is available through OpenRouter and Fireworks AI. As of June 2026, the managed cloud launches of Grok to AWS Bedrock and Azure AI Catalog applied specifically to Grok 4.3; Grok 4.20 access on those platforms is routed via third-party gateways. The model is closed-weight and proprietary: it cannot be self-hosted, fine-tuned, or deployed in air-gapped environments. xAI's earlier models, Grok 1 (Apache 2.0) and Grok 2 (community license), were open-weight, but starting with Grok 3, xAI moved to closed weights for its flagship series.

On safety and alignment, xAI applies a combination of reinforcement learning from human feedback and supervised fine-tuning with a stated design philosophy of being maximally truth-seeking. The published model cards covering Grok 4 and Grok 4.1 documented safety evaluations across abuse potential, concerning propensities, and dual-use capabilities. No separate model card was published for Grok 4.20 as of June 2026; xAI states safety methodology is continuous across the 4.x family. The model's default safety posture is configurable via system prompt, giving operators more control over refusal behavior than models with fixed alignment policies. Grok 4.20 is generally more permissive than Claude Opus 4.8 or GPT-5 on edge-case requests by default, which suits some use cases and creates risk for others.

Grok 4.20 is best suited for teams working with document sets exceeding 200K tokens, where its 2 million token context window is a hard competitive advantage. Its GPQA Diamond score makes it a strong choice for scientific reasoning pipelines in chemistry, biology, and physics. The 4-agent architecture benefits long-form content generation and multi-step research tasks requiring internal fact-checking. Its output speed serves latency-tolerant batch analysis well. Teams that should look elsewhere include those building real-time voice applications (no audio I/O, a 15-second time-to-first-token on reasoning variants), organizations requiring on-premise or air-gapped deployment, and enterprise teams needing strict content controls where GPT-5 or Claude Opus 4.8 are safer defaults.

The static training data cutoff is November 2024. xAI uses a mix of public web text, licensed datasets, and synthetic reasoning traces; a detailed training data breakdown has not been publicly disclosed. Real-time X platform data extends the effective knowledge boundary for current-events queries. API inputs are not used for model training by default; a zero-retention enterprise option is available on request. Default data retention is 30 days for safety monitoring, then deleted unless flagged. SOC 2 Type II certification status for the Grok 4.20 API tier had not been independently confirmed as of June 2026. EU AI Act classification applies under the general-purpose AI with systemic risk obligations category.

Grok 4.20 Beta launched February 17, 2026. Beta 2 on March 3, 2026, delivered five targeted fixes covering instruction following, hallucination reduction, LaTeX rendering, image search accuracy, and multi-image support. The reasoning API variant launched March 10, 2026. Full GA rollout followed March 18, 2026. Grok 4.3 became the new flagship on April 30, 2026, with Grok 4.20 remaining in service as the long-context-specialized option. Grok 5, targeting native audio and video multimodality, is on the public roadmap.

Pricing

$1.25 per 1M input tokens and $2.50 per 1M output tokens across the non-reasoning, reasoning, and multi-agent variants as of May 2026, cut from $2.00 and $6.00 per 1M at the February 2026 launch. Cached input runs $0.20 per 1M tokens (an 84% discount). Heavy mode's larger agent count uses the same rate card but consumes far more tokens per task, since every agent in the council generates its own reasoning trace.

What a real job costs

Budgets: 20-page PDF = 30k in / 1k out · Support reply = 2k in / 300 out · Coding agent run = 200k in / 20k out. Computed from the vendor's per-token prices at render time; cached-input discounts are not applied.

Key Features

• 4-Agent Inference Council: Four specialized agents (coordinator, research, math/code, and synthesis) debate and cross-check every answer in parallel, cutting the hallucination rate from about 12% to 4.2% in xAI's internal testing.
• 2 Million Token Context (Multi-Agent Variant): The multi-agent variant processes up to 2 million tokens per call, well beyond the 1 million ceiling of the single-model reasoning and non-reasoning variants.
• Faster Output Than Peer Frontier Models: Delivers output at a rate 3 to 4 times quicker than comparable frontier models, according to Artificial Analysis's speed benchmarks published in April 2026.
• Real-Time X Platform Data: A live search hook pulls current X platform posts at inference time, extending the model's effective knowledge past its static November 2024 training cutoff.
• Configurable Safety Posture: Operators can tighten or loosen refusal behavior through the system prompt, unlike models that ship with one fixed alignment policy.

Pros

• The 2 million token context window on the multi-agent variant removes chunking entirely for document sets other flagship models can't fit in one call.
• Output keeps pace on batch workloads that leave other frontier models waiting, per Artificial Analysis's 2026 speed testing.
• Priced well below Claude Opus 4.8 and GPT-5 on both input and output tokens, per xAI's published May 2026 rate card.

Cons

• No native audio or video input, so voice and multimodal front-ends need a separate preprocessing model bolted on.
• Running Heavy mode's 16 parallel agents multiplies token consumption 8 to 12 times per task, an expensive choice for short, high-volume workloads.
• Closed weights and API-only access: no self-hosting, fine-tuning, or Bedrock/Azure managed deployment for the 4.20 tier as of June 2026.

Benchmarks

• MMLU-Pro: 86.6% vendor-reported · 20 Mar 2026 — A harder version of the 57-subject knowledge exam, % correct.
• AIME 2025: 95% vendor-reported · 20 Mar 2026 — Competition-level maths problems from the 2025 exam, % solved.
• LMArena Elo: 1520 vendor-reported · 20 Mar 2026 — Rating from blind human votes on which answer is better.
• GPQA Diamond: 88.9% vendor-reported · 20 Mar 2026 — PhD-level science questions that are hard to search for, % correct.
• SWE-bench Verified: 75% vendor-reported · 20 Mar 2026 — Real GitHub issues fixed end to end, % solved.
• AA blended price: $1.88/M cited: Artificial Analysis · 01 Apr 2026 — Price per 1M tokens at a 3:1 input to output blend, as listed by Artificial Analysis.
• Output speed: 265 tok/s cited: Artificial Analysis · 01 Apr 2026 — Median tokens written per second as measured by Artificial Analysis.

A benchmark is an exam, not the job. Scores transfer unevenly between tasks, so weigh the one closest to your workload and read every figure with its source.

Visit Grok 4.20 Official Page