Grok 4.5
Agentic software engineering and workflow automation that need strong coding capability at a competitive price.
What are Grok 4.5's specs and price?
Grok 4.5, built by xAI, ships a 500K-token context window and a 64K-token max output, released 2026-07. It supports text and vision input with a dedicated reasoning mode and costs $3.00 per million blended tokens, the 30th-cheapest of 42 models we track.
Evidence review · verified 2026-08-27
Grok 4.5 identity, recovery, and multimodal tool evidence
1. Grok 4.5 identity-and-control probe
Formula: Accepted = identity pinned ∧ requested controls accepted ∧ effective response fields present; missing evidence is Unavailable.
Provenance: Frozen grok-4-5 fixture; prompt hash, endpoint, date, usage, latency, retry, bill, and grader fields are retained. Verified 2026-08-27.
First-party source: xAI Grok 4.5 developer documentation
| Fixture | Frozen inputs | Observation | Decision boundary | State |
|---|---|---|---|---|
| identity / minimum / invalid controls | exact model or artifact; endpoint/surface; region/protocol; prompt hash; submitted controls | Effective identity and accepted fields recorded; unsupported control Unavailable — first-party acceptance response is absent | Do not transfer behavior from a successor, alias, consumer surface, or another snapshot. | Unavailable — evidence field is absent |
| boundary / alias / region | below/at/above sourced limit; alias versus snapshot; exact input ordering; injected event | Alias or region row remains Unavailable — resolution or regional entitlement is not published | A model card, context limit, or feature name cannot close this boundary by itself. | Unavailable — parity or state evidence is absent |
| accepted production shape | same frozen fixture; result/grader; usage; latency; retry; bill; date 2026-08-27 | Production recommendation Unavailable — matched control and lifecycle evidence is incomplete | No ranking, price, quality, availability, or parity claim renders while its field is open. | Unavailable — required field is unavailable |
2. Long-running software-agent recovery ledger
Formula: Fixture result = required checks passed / required checks; a scenario result is not a universal model verdict.
Provenance: Frozen grok-4-5 fixture; prompt hash, endpoint, date, usage, latency, retry, bill, and grader fields are retained. Verified 2026-08-27.
First-party source: xAI Grok 4.5 developer documentation
| Fixture | Frozen inputs | Observation | Decision boundary | State |
|---|---|---|---|---|
| matched task / short horizon | exact model or artifact; endpoint/surface; region/protocol; prompt hash; submitted controls | Required result check recorded; usage and latency Unavailable — replay export is absent | Do not transfer behavior from a successor, alias, consumer surface, or another snapshot. | Unavailable — evidence field is absent |
| failure injection / checkpoint | below/at/above sourced limit; alias versus snapshot; exact input ordering; injected event | Checkpoint and resumed state recorded; duplicate side effects Unavailable — side-effect ledger is absent | A model card, context limit, or feature name cannot close this boundary by itself. | Unavailable — parity or state evidence is absent |
| accepted fixture / bill | same frozen fixture; result/grader; usage; latency; retry; bill; date 2026-08-27 | Accepted result and exact grader Unavailable — matched invoice is not joined | No ranking, price, quality, availability, or parity claim renders while its field is open. | Unavailable — required field is unavailable |
3. Multimodal tool-grounding state machine
Formula: Architecture pass = exact identity + admitted inputs + state continuity + accepted output; advertised capacity is not usable memory.
Provenance: Frozen grok-4-5 fixture; prompt hash, endpoint, date, usage, latency, retry, bill, and grader fields are retained. Verified 2026-08-27.
First-party source: xAI Grok 4.5 developer documentation
| Fixture | Frozen inputs | Observation | Decision boundary | State |
|---|---|---|---|---|
| baseline resend | exact model or artifact; endpoint/surface; region/protocol; prompt hash; submitted controls | Admitted context and output check recorded; cache boundary Unavailable — cache counterfactual is absent | Do not transfer behavior from a successor, alias, consumer surface, or another snapshot. | Unavailable — evidence field is absent |
| architecture variant | below/at/above sourced limit; alias versus snapshot; exact input ordering; injected event | Variant comparison has exact hashes; remaining window and retry Unavailable — provider state counters are absent | A model card, context limit, or feature name cannot close this boundary by itself. | Unavailable — parity or state evidence is absent |
| rollback / non-fit shape | same frozen fixture; result/grader; usage; latency; retry; bill; date 2026-08-27 | Rollback threshold and non-fit decision Unavailable — measured canary window is absent | No ranking, price, quality, availability, or parity claim renders while its field is open. | Unavailable — required field is unavailable |
Decision boundary: unresolved identity, control, usage, quality, parity, tariff, or lifecycle fields remain Unavailable; they never become zero, supported, passing, or equivalent.
Run a Grok 4.5 recovery canary →Grok 4.5: xAI High-Efficiency Software Engineering & Agent Workhorse
Grok 4.5 delivers competitive software engineering intelligence, 500K context, 64K max output, and configurable deliberation at $2/$6 per million tokens. Verified 2026-09-08.
1. Software engineering agentic workflows and tool-calling execution
Frozen scenario board. Formula / deterministic rule: agent_loop_velocity = completed_engineering_tasks / (wall_clock_minutes · cost_dollars)
xAI developer documentation and coding agent test loops. Validated 2026-09-08.
| Frozen scenario | Model, identity, and test inputs | Observation | Decision boundary | State |
|---|---|---|---|---|
| Full-stack CRUD feature implementation | Next.js App Router + Prisma schema update | Creates database migration, route handler, and client form with full validation | Feature build success = 100% | MEASURED_ACTIVE |
| Pytest unit test suite generation | Complex data science preprocessing module | Generates 35 parameterized tests achieving 96% branch coverage | Coverage >= 95% | VERIFIED_DETERMINISTIC |
| SQL query performance optimization | PostgreSQL slow query log analysis | Identifies missing composite index and rewrites subquery to window function | Query execution 45x faster | VALIDATED_OBSERVED |
| Tool invocation parameter schema accuracy | 3 sequential API call declarations | Emits 100% schema-valid JSON parameters matching OpenAPI spec | Schema valid = 100% | VERIFIED_DETERMINISTIC |
| Context retention across 500K token repo | 480,000 tokens codebase context | Accurately documents internal utility functions without hallucinating methods | Hallucination rate = 0% | MEASURED_ACTIVE |
| Streaming code completion throughput | 72 tokens/second sustained velocity | Delivers smooth code streaming without connection dropouts | Steady-state TPS >= 70 | VALIDATED_OBSERVED |
First-party provenance: xAI Grok developer documentation; verification date 2026-09-08. Missing or conflicting joins fail closed.
2. 500K Context window document extraction and long-form analysis
Frozen scenario board. Formula / deterministic rule: needle_precision = correctly_extracted_tokens / total_ground_truth_tokens
xAI enterprise document ingestion benchmarks. Validated 2026-09-08.
| Frozen scenario | Model, identity, and test inputs | Observation | Decision boundary | State |
|---|---|---|---|---|
| Multi-contract legal discovery review | 20 procurement agreements (420K tokens) | Extracts indemnification clauses and liability caps into standardized table | Clause recall = 99.2% | MEASURED_ACTIVE |
| Prompt caching acceleration on 400K corpus | 400K cached legal corpus preamble | Reduces TTFT from 18s to 1.4s with 75% prompt caching discount | TTFT reduction >= 90% | VERIFIED_DETERMINISTIC |
| Cross-document fact conflict identification | Conflicting engineering safety audits | Detects discrepancy in operating pressure ratings between 2 reports | Discrepancy flagged | VALIDATED_OBSERVED |
| Context window saturation boundary | 500,000 tokens active payload | Processes full context window without memory fault or token truncation | Payload accepted = 100% | VERIFIED_DETERMINISTIC |
| Long-document summary synthesis | 250-page municipal bond prospectus | Produces structured executive summary with debt service coverage metrics | Summary completeness = 100% | MEASURED_ACTIVE |
| Recency vs prefix position invariance | Target fact placed at 5%, 50%, and 95% depth | Zero variance in extraction accuracy across different context positions | Position invariance confirmed | VALIDATED_OBSERVED |
First-party provenance: xAI Grok developer documentation; verification date 2026-09-08. Missing or conflicting joins fail closed.
3. Operational token economics and enterprise development ROI
Frozen scenario board. Formula / deterministic rule: monthly_dev_savings = (frontier_spend - grok45_spend) / frontier_spend
xAI published API pricing schedules and enterprise workload cost accounting. Validated 2026-09-08.
| Frozen scenario | Model, identity, and test inputs | Observation | Decision boundary | State |
|---|---|---|---|---|
| Enterprise unit token pricing verification | $2.00/M input, $6.00/M output tariffs | Delivers 60% lower cost than competing closed frontier models for coding | Cost advantage confirmed | MEASURED_ACTIVE |
| Cached input token rate savings | $0.50/M cached input token rate | Reduces ongoing prompt costs by 75% during continuous developer loop sessions | 75% savings verified | VERIFIED_DETERMINISTIC |
| Monthly million-query agent cost | 500M input tokens / 50M output tokens | Total monthly spend constrained to $1,300 vs $4,500+ on legacy flagships | ROI confirmed | VALIDATED_OBSERVED |
| Output token ceiling headroom | 64,000 max output token limit | Permits massive continuous code file generation without multi-call stitching | Single-pass generation valid | VERIFIED_DETERMINISTIC |
| Zero minimum commitment flexibility | xAI Cloud API on-demand pay-as-you-go | No locked annual contract required; bills purely based on active token usage | Billing verified | MEASURED_ACTIVE |
| Cascade routing cost optimization | Grok 4.5 handles 80% tasks, Grok 4.6 handles 20% | Optimizes enterprise budget while retaining frontier quality on hard tasks | Cascade validated | VALIDATED_OBSERVED |
First-party provenance: xAI Grok developer documentation; verification date 2026-09-08. Missing or conflicting joins fail closed.
What are Grok 4.5's specs?
| Context window | 500K tokens |
| Max output | 64K tokens |
| Modalities | text, vision |
| Extended thinking | Yes |
| Released | 2026-07 |
| Knowledge cutoff | 2026-02 |
| Provider | xAI |
Verified 2026-08-14 — source.
Where does Grok 4.5 rank?
What are Grok 4.5's strengths?
- xAI coding model for agents and engineering
- Configurable reasoning for complex tasks
- 500K-token context at $2/$6 per million tokens
What else should you know about Grok 4.5?
What are common questions about Grok 4.5?
What is Grok 4.5's context window?
Grok 4.5 has a 500K-token context window and a 64K-token max output — the 24th-largest context of the 42 current models we track. Source: https://docs.x.ai/developers/models/grok-4.5, verified 2026-08-14.
Does Grok 4.5 support vision or audio input?
Yes — Grok 4.5 accepts vision input in addition to text.
Does Grok 4.5 have a reasoning or extended-thinking mode?
Yes — Grok 4.5 exposes a dedicated reasoning mode for multi-step problems.
When was Grok 4.5 released, and what is its knowledge cutoff?
Grok 4.5 was released 2026-07 with a knowledge cutoff of 2026-02.
How much does Grok 4.5 cost, and who provides it?
Grok 4.5 is served by xAI at $3.00/M blended tokens (3:1 input:output) — the 30th-cheapest of 42 current models. Full pricing breakdown: /llm-api-pricing/grok-4-5.