Grok 4.20 Reasoning API Pricing: Dedicated Thinking Tokens at Scale
Detailed xAI Grok 4.20 Reasoning API pricing ($2.00/M input, $6.00/M output), dedicated reasoning token budgets, 2M context window, and enterprise token costs.
Full specs, context window and API limits →How much does Grok-4.20 Reasoning cost per million tokens?
xAI Grok 4.20 Reasoning charges $2.00 per million input tokens and $6.00 per million output tokens ($3.00/M blended at 3:1). Features dedicated reasoning tokens, 2M context, and live data search. Verified 2026-09-08.
Verified 2026-09-07 — source
How much does Grok-4.20 Reasoning cost per 1,000 requests?
Computed from generated token pricing. Each row assumes the listed input and output tokens per request; this model has no measured verbosity factor, so the unadjusted output estimate is shown.
| Request shape | Input tokens | Output tokens | Cost / 1,000 requests |
|---|---|---|---|
| Short | 100 | 50 | $0.5000 |
| Medium | 1,000 | 500 | $5.0000 |
| Long | 4,000 | 2,000 | $20.0000 |
Formula: ((input price × input tokens) + (output price × output tokens × verbosity factor)) ÷ 1,000,000 × 1,000. Assumptions: short 100/50, medium 1,000/500, long 4,000/2,000 input/output tokens per request. Unadjusted — no measured verbosity factor is available.
Three model-specific pricing decisions
The reasoning variant uses the same listed token rate as its paired variant in the current registry. Its premium is therefore exposed as output expansion, not invented as a hidden per-token surcharge.
1. Equal-rate audit and reasoning-token expansion
| Shape | Listed rate | Reasoning output assumption | 100K cost |
|---|---|---|---|
| Chat | $2.00 in / $6.00 out | 500 visible output | $700.00 |
| Expanded reasoning | Same listed rate | 2× output tokens | $1000.00 |
| Heavy reasoning | Same listed rate | 4× output tokens | $1600.00 |
2. Accepted-answer uplift and retry crossover against Grok 4.3
| Fixed workload (input / output) | Grok-4.20 Reasoning | Grok 4.3 | Decision boundary |
|---|---|---|---|
| Reasoning coding · 4,000 / 1,000 | $1400.00 | $750.00 | 15% accepted-result uplift required to justify the higher bill |
| Long agent · 16,000 / 2,000 | $4400.00 | $2500.00 | 20% accepted-result uplift required to justify the higher bill |
Formula: 100,000 × (input tokens × input $/M + output tokens × output $/M) ÷ 1,000,000. Uplift is a decision threshold, not a measured quality claim.
3. 1M-context operational-risk table
| Dimension | Dated evidence | Cost/decision treatment |
|---|---|---|
| Token price | $2.00 / $6.00 per M | Sourced registry input |
| Latency | 52 tokens/sec | Do not convert speed to dollars |
| Context fit | 1M context claim unavailable in pricing record | Check model specification owner |
| Cache / batch | Unavailable | No zero-cost assumption |
| SLA | Unavailable | No SLA claim |
Verified 2026-04-06. “Unavailable” means no compatible dated evidence was found; it is never treated as zero. First-party price source · Run this exact scenario.
All three decisions are server-rendered for Grok-4.20 Reasoning; fixed inputs, formulas, dated sources, and unavailable states are intentionally visible.
Exact-model pricing guide · verified 2026-09-07
Exact model boundary: xAI grok-4.20-0309-reasoning (slug grok-4-20-0309-reasoning). First-party provider pricing and API documentation remain fact owners.
Reasoning token budget and monthly expenditure matrix
Frozen scenario board. Formula / deterministic rule: monthly_spend = calls * ((in_tokens * 2.00 + out_tokens * 6.00) / 1M) Boundary: Owns reasoning token expenditure modeling for Grok 4.20 Reasoning.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| 5K mathematical proofs | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=5K mathematical proofs; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 5K mathematical proofs is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 25K algorithmic coding challenges | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=25K algorithmic coding challenges; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 25K algorithmic coding challenges is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 50K architectural design reviews | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=50K architectural design reviews; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 50K architectural design reviews is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| prompt caching active (75% read discount) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=prompt caching active (75% read discount); workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — prompt caching active (75% read discount) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| batch processing active (50% both) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=batch processing active (50% both); workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — batch processing active (50% both) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unresolved pricing currency | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=unresolved pricing currency; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — unresolved pricing currency has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: xAI Grok developer pricing; verification date 2026-09-07. Missing or conflicting joins fail closed.
Context window saturation (2M scale) capability gate
Frozen scenario board. Formula / deterministic rule: eligible = (context <= 2M ? Eligible : Excluded); extreme long-context reasoning Boundary: Owns context scaling and capability gating for large inputs.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| small library codebase (50K tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=small library codebase (50K tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — small library codebase (50K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| medium application service (250K tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=medium application service (250K tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — medium application service (250K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| large monolithic codebase (500K tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=large monolithic codebase (500K tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — large monolithic codebase (500K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| enterprise monorepo (1M tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=enterprise monorepo (1M tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — enterprise monorepo (1M tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 2M full context saturation test | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=2M full context saturation test; context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 2M full context saturation test is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unsupported archive payload format | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=unsupported archive payload format; context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — unsupported archive payload format has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: xAI Grok documentation; verification date 2026-09-07. Missing or conflicting joins fail closed.
Deliberation budget control and deep mathematical proof accuracy
Frozen scenario board. Formula / deterministic rule: proof_roi = saved_engineering_hours - ((in_tokens * 2.00 + out_tokens * 6.00) / 1M) Boundary: Owns reasoning depth modeling and return-on-investment calculations.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| formal mathematical verification | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=formal mathematical verification; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — formal mathematical verification is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| cryptographic protocol validation | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=cryptographic protocol validation; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — cryptographic protocol validation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| distributed consensus protocol proof | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=distributed consensus protocol proof; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — distributed consensus protocol proof is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| optimizing compiler intermediate representation | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=optimizing compiler intermediate representation; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — optimizing compiler intermediate representation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| reasoning token budget exhaustion | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=reasoning token budget exhaustion; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — reasoning token budget exhaustion has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
| unresolved proof specification | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=unresolved proof specification; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — unresolved proof specification has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: xAI Grok documentation; verification date 2026-09-07. Missing or conflicting joins fail closed.
Method and limitations: formulas are deterministic; observed and assumed inputs are labeled; no missing provider, host, account, region, realm, alias, snapshot, revision, weight, artifact, control, tool, modality, workload, rate-period, timestamp, or result is transferred. Run this scenario →
Verified Model Pricing Intelligence•Audit date: 2026-09-08
Grok 4.20 Reasoning API Pricing: Dedicated Thinking Tokens at Scale
xAI Grok 4.20 Reasoning charges $2.00 per million input tokens and $6.00 per million output tokens ($3.00/M blended at 3:1). Features dedicated reasoning tokens, 2M context, and live data search. Verified 2026-09-08.
Reasoning token budget and monthly expenditure matrix
Frozen scenario board. Formula / deterministic rule: monthly_spend = calls * ((in_tokens * 2.00 + out_tokens * 6.00) / 1M) Boundary: Owns reasoning token expenditure modeling for Grok 4.20 Reasoning.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| 5K mathematical proofs | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=5K mathematical proofs; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 5K mathematical proofs is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 25K algorithmic coding challenges | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=25K algorithmic coding challenges; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 25K algorithmic coding challenges is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 50K architectural design reviews | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=50K architectural design reviews; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 50K architectural design reviews is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| prompt caching active (75% read discount) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=prompt caching active (75% read discount); workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — prompt caching active (75% read discount) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| batch processing active (50% both) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=batch processing active (50% both); workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — batch processing active (50% both) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unresolved pricing currency | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=unresolved pricing currency; workload; prompt tokens; completion tokens; standard spend; cached spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — unresolved pricing currency has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: xAI Grok developer pricing; verification date 2026-09-08. Missing or conflicting joins fail closed.
Context window saturation (2M scale) capability gate
Frozen scenario board. Formula / deterministic rule: eligible = (context <= 2M ? Eligible : Excluded); extreme long-context reasoning Boundary: Owns context scaling and capability gating for large inputs.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| small library codebase (50K tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=small library codebase (50K tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — small library codebase (50K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| medium application service (250K tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=medium application service (250K tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — medium application service (250K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| large monolithic codebase (500K tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=large monolithic codebase (500K tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — large monolithic codebase (500K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| enterprise monorepo (1M tokens) | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=enterprise monorepo (1M tokens); context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — enterprise monorepo (1M tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 2M full context saturation test | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=2M full context saturation test; context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 2M full context saturation test is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unsupported archive payload format | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=unsupported archive payload format; context size; codebase eligible; prompt cost; completion cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — unsupported archive payload format has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: xAI Grok documentation; verification date 2026-09-08. Missing or conflicting joins fail closed.
Deliberation budget control and deep mathematical proof accuracy
Frozen scenario board. Formula / deterministic rule: proof_roi = saved_engineering_hours - ((in_tokens * 2.00 + out_tokens * 6.00) / 1M) Boundary: Owns reasoning depth modeling and return-on-investment calculations.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| formal mathematical verification | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=formal mathematical verification; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — formal mathematical verification is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| cryptographic protocol validation | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=cryptographic protocol validation; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — cryptographic protocol validation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| distributed consensus protocol proof | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=distributed consensus protocol proof; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — distributed consensus protocol proof is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| optimizing compiler intermediate representation | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=optimizing compiler intermediate representation; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — optimizing compiler intermediate representation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| reasoning token budget exhaustion | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=reasoning token budget exhaustion; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — reasoning token budget exhaustion has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
| unresolved proof specification | model=grok-4.20-0309-reasoning; slug=grok-4-20-0309-reasoning; provider=xAI; scenario=unresolved proof specification; reasoning problem; thinking depth budget; token cost; net economic ROI; recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — unresolved proof specification has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: xAI Grok documentation; verification date 2026-09-08. Missing or conflicting joins fail closed.
Exact-Model Pricing Evidence•Verified 2026-04-06; revalidation required before current claims.
Grok 4.20 Reasoning pricing evidence
Source-backed dated rate shape: input=$2.000000/M; output=$6.000000/M. Grok 4.20 reasoning-variant token accounting only; xAI policy, non-reasoning speed claims, and math verdicts retain their owners. Missing evidence, failed revalidation, and unsupported mechanics render Unavailable.
Module 1 of 3: Deliberation-token spend ladder
Novel contribution boundary: Owns dated reasoning-variant arithmetic for fixed deliberation-shaped outputs; reasoning quality and live-search behavior are not asserted. Formula / deterministic rule: spend = requests × (inputTokens × inputRate + outputTokens × outputRate) / 1,000
| Scenario | Exact model, provider, and fixed inputs | Result | State |
|---|---|---|---|
| 5K compact proof requests | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=requests=5,000; inputTokens=2,000; outputTokens=1,000; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| 25K algorithm-design requests | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=requests=25,000; inputTokens=6,000; outputTokens=3,000; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| 50K deep deliberation requests | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=requests=50,000; inputTokens=12,000; outputTokens=8,000; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
First-party registry provenance: https://docs.x.ai/docs/overview#pricing; registry provider xai; verified 2026-04-06; freshness gate=2026-09-14. Revalidate before any current-price claim.
Module 2 of 3: Reasoning→non-reasoning accepted-result crossover
Novel contribution boundary: Owns dated variant arithmetic with explicit retry input; accepted-result quality and reasoning overhead are not sourced. Formula / deterministic rule: requiredAcceptedRate = non-reasoning attempt-plus-retry cost / reasoning attempt cost; observed acceptance = Unavailable
| Scenario | Exact model, provider, and fixed inputs | Result | State |
|---|---|---|---|
| Short proof; 10% retry share; 500 retry-input tokens | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=requests=1; inputTokens=2,000; outputTokens=1,000; retryInputTokens=500; retryShare=10%; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| Medium algorithm review; 20% retry share; 1,000 retry-input tokens | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=requests=1; inputTokens=6,000; outputTokens=3,000; retryInputTokens=1,000; retryShare=20%; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| Long architecture proof; 30% retry share; 2,000 retry-input tokens | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=requests=1; inputTokens=12,000; outputTokens=8,000; retryInputTokens=2,000; retryShare=30%; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
First-party registry provenance: https://docs.x.ai/docs/overview#pricing; registry provider xai; verified 2026-04-06; freshness gate=2026-09-14. Revalidate before any current-price claim.
Module 3 of 3: Live-search and context evidence map
Novel contribution boundary: Owns exact-model registry evidence only; live-search fees and context treatment are not inferred. Formula / deterministic rule: evidence = exact-model registry field when present; undocumented fee or limit = Unavailable
| Scenario | Exact model, provider, and fixed inputs | Result | State |
|---|---|---|---|
| Live-search fee or eligibility | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=evidenceField=live-search; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| Context-window pricing treatment | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=evidenceField=context; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| Reasoning-token accounting field | model=grok-4.20-0309-reasoning; provider=xai; registryRates=input=$2.000000/M; output=$6.000000/M; comparisonModel=grok-4.20-0309-non-reasoning; fixedInputs=evidenceField=modality; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-04-06, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
First-party registry provenance: https://docs.x.ai/docs/overview#pricing; registry provider xai; verified 2026-04-06; freshness gate=2026-09-14. Revalidate before any current-price claim.
How fast is Grok-4.20 Reasoning?
Tokens / sec
52
TTFT
540 ms
Rank
#27 of 31
$ / M ÷ t/s
$0.06
Measured with 5 runs on a fixed prompt — see the full methodology.
How much does Grok-4.20 Reasoning cost at scale?
| Tokens / month | Est. cost (blended 3:1) |
|---|---|
| 100,000 | $0.30 |
| 1,000,000 | $3.00 |
| 10,000,000 | $30.00 |
| 100,000,000 | $300.00 |
How does Grok-4.20 Reasoning compare with other models?
Grok-3 Mini — $0.26/MGrok 4.3 — $1.56/MGrok-3 — $2.50/MGrok-4.20 — $3.00/MGrok 4.6 — $3.00/MGrok-4.20 — $3.00/MGrok 4.6 — $3.00/MGrok 4.5 — $3.00/M
What is Grok-4.20 Reasoning best for?
What should you explore next for Grok-4.20 Reasoning?
Looking for a cheaper option?
Muse Spark 1.3 Contributor is 95.8% cheaper — a config migration. See all 8 alternatives to Grok-4.20 Reasoning →
Which Grok-4.20 Reasoning head-to-head comparisons are available?
What are common questions about Grok-4.20 Reasoning?
Is Grok-4.20 Reasoning cheaper than Grok-4.20?
Grok-4.20 Reasoning costs $3.00/M blended tokens, Grok-4.20 costs $3.00/M — Grok-4.20 is cheaper.
How much does 1 million tokens cost with Grok-4.20 Reasoning?
At a 3:1 input:output ratio, 1 million blended tokens costs approximately $3.00. Pure input costs $2.00/M; pure output costs $6.00/M.
What does Grok-4.20 Reasoning cost at high volume?
At 100 million blended tokens a month, Grok-4.20 Reasoning costs approximately $300.00. See the cost-at-scale table below for other volumes.