Compare Grok-4.20 Reasoning and Grok 4.3: Analytical Reasoning vs High-Speed Streaming
Grok-4.20 Reasoning vs Grok 4.3: which should I use?
Move from Grok 4.20 0309 reasoning to Grok 4.3 only after requested and effective IDs, mode, cluster, response controls, and replay safety are proven for the workload. Keep the dated endpoint as fallback. Verified 2026-09-02; ambiguous aliases and unsafe writes remain Blocked.
Verified 2026-09-02
Where can you find price, speed, and task evidence for Grok-4.20 Reasoning and Grok 4.3?
Which tasks fit Grok-4.20 Reasoning and Grok 4.3?
Best-fit task signals from the published model strengths; this is not a substitute for a controlled benchmark.
| Fact | Grok-4.20 Reasoning | Grok 4.3 |
|---|---|---|
| Best fit | Long-document analysis and problems that benefit from explicit reasoning. | General-purpose work where speed and a huge context window both matter. |
| Reasoning mode | Available | Available |
What does a Coding Agent workload cost with Grok-4.20 Reasoning and Grok 4.3?
Modeled for Coding Agent: 20,000 input + 2,000 output tokens per task, 1 turn. This is a workload model, not a provider quote.
| Fact | Grok-4.20 Reasoning | Grok 4.3 |
|---|---|---|
| Coding Agent / task | $0.05 (modeled) | $0.03 (modeled) (winner) |
| Input / output rate | $2.00 / $6.00 per M | $1.25 / $2.50 per M |
How fast are Grok-4.20 Reasoning and Grok 4.3?
Only non-estimated benchmark results are shown as measured.
| Fact | Grok-4.20 Reasoning | Grok 4.3 |
|---|---|---|
| Measured throughput | 52 tokens/s | 98 tokens/s (winner) |
| Time to first token | 540 ms | 320 ms |
How compatible are Grok-4.20 Reasoning and Grok 4.3 with APIs?
Provider-level wire-format and SDK facts; model-specific parameter differences may still apply.
| Fact | Grok-4.20 Reasoning | Grok 4.3 |
|---|---|---|
| OpenAI SDK | Usable | Usable |
| Request shape | Full OpenAI chat/completions compatibility — point the openai SDK at api.x.ai and it works unmodified. | Full OpenAI chat/completions compatibility — point the openai SDK at api.x.ai and it works unmodified. |
| Streaming | SSE (OpenAI delta) | SSE (OpenAI delta) |
What are the privacy and retention policies for Grok-4.20 Reasoning and Grok 4.3?
Provider policy and residency facts from the verified provider registry. A missing retention policy is not treated as a guarantee.
| Fact | Grok-4.20 Reasoning | Grok 4.3 |
|---|---|---|
| Provider says API data trains models | Unavailable | Unavailable |
| Published retention period | Unavailable | Unavailable |
| Data residency | Unavailable | Unavailable |
How much effort does it take to migrate between Grok-4.20 Reasoning and Grok 4.3?
Direction is from each displayed model to the other. Effort is derived from provider compatibility and published parameter maps.
| Fact | Grok-4.20 Reasoning | Grok 4.3 |
|---|---|---|
| Move Grok-4.20 Reasoning → Grok 4.3 | drop-in; 0 breaking parameter differences | Target: Grok 4.3 |
| Move Grok 4.3 → Grok-4.20 Reasoning | Source: Grok 4.3 | drop-in; 0 breaking parameter differences |
| Why | Same provider; only the model identifier changes. | Same provider; only the model identifier changes. |
Batch 54 · exact pair cutover and operating-tier contributions. Surface verification: 2026-09-02. These exactly three boards are server-rendered from frozen scenario fixtures; assumptions and unavailable joins are not observations.
Exact pair boundary: xAI; dated reasoning snapshot and successor exact ID/mode/cluster must agree; aliases do not inherit identity. Requested models are grok-4.20-0309-reasoning and grok-4.3. Provider, rate, pricing, and task pages remain fact owners.
Dated-to-successor endpoint identity resolver
Frozen Batch 54 scenario board. Formula / deterministic rule: identity_ok = requested ID resolves to effective ID ∧ snapshot/mode/cluster/lifecycle agree; ambiguous => block Boundary: Exact Grok 4.20 reasoning-to-4.3 identity only; non-reasoning and later-version owners retain their scopes.
| Frozen scenario / field ID | Pair, host, surface, and evidence fields | Result | State |
|---|---|---|---|
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r14.20 reasoning ID | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; requested=dated reasoning ID; snapshot=0309; mode=reasoning; cluster=required; response ID=required; lifecycle/source date=dated; pin decision=required; evidence=2026-09-02 | Pin the dated reasoning ID and admit it as baseline only when the response identity returns the same snapshot. | CONDITIONAL — baseline identity |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r24.20 latest alias | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; requested=4.20 latest; alias=latest; mode=reasoning; cluster=required; effective ID/snapshot=required; response mode=required; lifecycle=dated; evidence=2026-09-02 | Unavailable — latest alias may not resolve to the dated 0309 snapshot | BLOCKED — alias identity |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r34.3 exact ID | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; requested=4.3 exact; mode=required; cluster=required; response ID=required; snapshot/lifecycle=dated; rate period=required; evidence=2026-09-02 | Unavailable — successor response identity and rate period are not joined | UNAVAILABLE — successor identity |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r44.3 latest alias | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; requested=4.3 latest; alias=latest; mode=unknown; cluster=required; effective ID=unknown; mode=required; response ID=required; pin action=required; evidence=2026-09-02 | Do not compare a latest alias until it is pinned to an exact version and mode. | BLOCKED — alias unresolved |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r5gateway-prefixed alias | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=shared gateway / shared gateway; gateway / gateway; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; requested=grok/latest; provider/cluster=unknown; effective provider/ID/mode/cluster=unknown; response ID=required; lifecycle=unknown; evidence=2026-09-02 | Reject gateway evidence when effective provider, cluster, or model is unknown. | BLOCKED — gateway identity |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r6unknown cluster | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; requested IDs=exact; cluster=unknown; mode=reasoning; response IDs=known; cluster evidence=missing; pin/migrate decision=blocked; evidence=2026-09-02 | Identity confidence is insufficient without cluster; no migration verdict. | BLOCKED — cluster missing |
First-party provenance: xAI model documentation; verification date 2026-09-02. Missing or conflicting joins fail closed.
Reasoning-control and response-contract delta receipt
Frozen Batch 54 scenario board. Formula / deterministic rule: comparable = exact endpoint + control semantics + reasoning usage + tool/schema fields; same label is not equivalence Boundary: No equivalence is inferred from similarly named effort labels, and the non-reasoning sibling is outside this receipt.
| Frozen scenario / field ID | Pair, host, surface, and evidence fields | Result | State |
|---|---|---|---|
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r1default | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; control=default; mode=reasoning; cluster=exact; config hash=grok-control-001; reasoning token visibility=required; usage fields=required; evidence=2026-09-02 | Unavailable — default-control semantics and token accounting are not joined | UNAVAILABLE — default mapping |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r2low | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; control=low; mode=reasoning; endpoint exact; config hash=grok-control-002; supported/transformed/rejected state=required; result hash=missing; evidence=2026-09-02 | Unavailable — low-control mapping is not evidenced | UNAVAILABLE — control mapping |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r3high | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; control=high; mode=reasoning; endpoint exact; config hash=grok-control-003; reasoning/output tokens=required; acceptance rubric=required; evidence=2026-09-02 | Unavailable — matched high-control output is absent | UNAVAILABLE — high evidence |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r4max | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; control=max; mode=reasoning; endpoint exact; config hash=grok-control-004; response usage/tool fields=required; parser=versioned; evidence=2026-09-02 | Treat max as separate until xAI documents equivalent semantics and a parser probe passes. | NOT COMPARABLE — mapping unknown |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r5strict schema | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; control=default; schema=strict-005; schema hash=strict-005; parser/refusal fields=required; response ID=required; evidence=2026-09-02 | Unavailable — schema and refusal behavior are not paired | UNAVAILABLE — schema contract |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r6streamed tool call | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; control=high; tool=stream-006; event sequence hash=required; tool/result IDs=required; usage fields=required; parser change=required; evidence=2026-09-02 | Unavailable — stream/tool settlement is not observed | UNAVAILABLE — stream contract |
First-party provenance: xAI model documentation; verification date 2026-09-02. Missing or conflicting joins fail closed.
Cluster-safe migration and replay board
Frozen Batch 54 scenario board. Formula / deterministic rule: replay = failure classified ∧ result/schema valid ∧ idempotency policy satisfied ∧ attempt cap; unsafe write => block Boundary: Exact endpoint cutover with side-effect containment; provider policy and later Grok choices remain separate owners.
| Frozen scenario / field ID | Pair, host, surface, and evidence fields | Result | State |
|---|---|---|---|
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r1read-only answer | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; cohort=replay-001; clusters=exact; prompt/config/result hashes=paired; failure=transient; side_effect=none; attempt cap=2; evidence=2026-09-02 | Unavailable — paired replay observation is absent | HOLD — read-only probe |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r2provider search | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; cohort=replay-002; cluster=exact; query hash=required; source IDs/timestamps=required; tool trace=required; result hash=missing; evidence=2026-09-02 | Unavailable — search provenance and cluster replay are not joined | UNAVAILABLE — provenance |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r3strict extraction | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; cohort=replay-003; schema=strict; schema hash=paired; validator result=required; parser version=required; attempt cap=2; evidence=2026-09-02 | Unavailable — schema validation and retry evidence are absent | UNAVAILABLE — extraction replay |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r4idempotent write | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; cohort=replay-004; write=idempotent; idempotency key=required; checkpoint/result hashes=required; side_effect=contained; rollback=required; evidence=2026-09-02 | Replay only with the same idempotency key after settlement and target validation. | CONDITIONAL — idempotency proof |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r5partial streamed tool call | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; cohort=replay-005; stream=partial; event/tool IDs=required; completed effects=required; uncommitted state=required; fallback=4.20; evidence=2026-09-02 | Unavailable — partial event and effect state are unknown | BLOCKED — partial stream |
batch54-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r6non-idempotent write | pair=grok-4.20-0309-reasoning vs grok-4.3; provider/host/surface=api.x.ai / api.x.ai; xAI API / xAI API; requested IDs=grok-4.20-0309-reasoning,grok-4.3; effective IDs=grok-4.20-0309-reasoning,grok-4.3; cohort=replay-006; write=non-idempotent; idempotency=missing; side_effect=external; human owner=required; rollback artifact=required; evidence=2026-09-02 | Do not replay across clusters or endpoints; stop for human reconciliation. | BLOCKED — unsafe replay |
First-party provenance: xAI API limits documentation; verification date 2026-09-02. Missing or conflicting joins fail closed.
Method and limitations: formulas are deterministic; assumptions are labeled; no missing provider, host, account, region, realm, cluster, alias, snapshot, version, artifact, effort, tool, modality, benchmark, workload, rate-period, timestamp, or result is transferred. Run the grok-4-20-0309-reasoning-vs-grok-4-3 Batch 54 scenario →
Batch 69 · exact-pair decision contributions · verified 2026-09-08
Exact pair boundary: xAI; same provider, tier transition, $2.00/$6.00 vs $1.25/$2.50 token tariffs, 2M vs 1M context, and reasoning tokens must join. Requested models are grok-4.20-0309-reasoning and grok-4.3. Provider, rate, pricing, and task pages remain fact owners.
Token pricing and reasoning token expenditure matrix
Frozen Batch 69 scenario board. Formula / deterministic rule: monthly_spend = calls * ((in_tokens * rate_in + out_tokens * rate_out) / 1M) Boundary: Owns base token tariff comparisons and workload expenditure modeling.
| Frozen scenario / field ID | Pair, identity, host, surface, and evidence fields | Result | State |
|---|---|---|---|
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r115K deep reasoning mathematical analyses | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=15K deep reasoning mathematical analyses; workload; prompt tokens; completion tokens; Grok-4.20 Reasoning spend; Grok 4.3 spend; budget delta; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 15K deep reasoning mathematical analyses is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r250K live-data monitoring passes | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=50K live-data monitoring passes; workload; prompt tokens; completion tokens; Grok-4.20 Reasoning spend; Grok 4.3 spend; budget delta; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 50K live-data monitoring passes is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r3150K high-speed conversational sessions | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=150K high-speed conversational sessions; workload; prompt tokens; completion tokens; Grok-4.20 Reasoning spend; Grok 4.3 spend; budget delta; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 150K high-speed conversational sessions is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r4prompt caching active (50% read discount both) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=prompt caching active (50% read discount both); workload; prompt tokens; completion tokens; Grok-4.20 Reasoning spend; Grok 4.3 spend; budget delta; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — prompt caching active (50% read discount both) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r5batch processing active (50% both) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=batch processing active (50% both); workload; prompt tokens; completion tokens; Grok-4.20 Reasoning spend; Grok 4.3 spend; budget delta; 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 |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m1-r6unresolved pricing currency | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=unresolved pricing currency; workload; prompt tokens; completion tokens; Grok-4.20 Reasoning spend; Grok 4.3 spend; budget delta; 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 API pricing documentation; verification date 2026-09-08. Missing or conflicting joins fail closed.
Context window saturation (2M vs 1M) capability gate
Frozen Batch 69 scenario board. Formula / deterministic rule: eligible = (context <= 1M ? Both : (context <= 2M ? Grok4_20 : Excluded)) Boundary: Owns context scaling and capability gating for massive document repositories.
| Frozen scenario / field ID | Pair, identity, host, surface, and evidence fields | Result | State |
|---|---|---|---|
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r1medium legal dossier (150K tokens) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=medium legal dossier (150K tokens); context size; Grok-4.20 eligible; Grok 4.3 eligible; Grok-4.20 cost; Grok 4.3 cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — medium legal dossier (150K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r2dense technical archive (500K tokens) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=dense technical archive (500K tokens); context size; Grok-4.20 eligible; Grok 4.3 eligible; Grok-4.20 cost; Grok 4.3 cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — dense technical archive (500K tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r3maximum Grok 4.3 context (1M boundary) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=maximum Grok 4.3 context (1M boundary); context size; Grok-4.20 eligible; Grok 4.3 eligible; Grok-4.20 cost; Grok 4.3 cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — maximum Grok 4.3 context (1M boundary) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r4massive corporate data lake (1.5M tokens) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=massive corporate data lake (1.5M tokens); context size; Grok-4.20 eligible; Grok 4.3 eligible; Grok-4.20 cost; Grok 4.3 cost; context gate verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — massive corporate data lake (1.5M tokens) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r52M full saturation audit | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=2M full saturation audit; context size; Grok-4.20 eligible; Grok 4.3 eligible; Grok-4.20 cost; Grok 4.3 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 saturation audit is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m2-r6unsupported archive payload format | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=unsupported archive payload format; context size; Grok-4.20 eligible; Grok 4.3 eligible; Grok-4.20 cost; Grok 4.3 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.
Live X search retrieval grounding and latency trade-off
Frozen Batch 69 scenario board. Formula / deterministic rule: turnaround = ttft + live_x_search_latency + (tokens_out / tps) Boundary: Owns real-time live data search integration and turn turnaround speed.
| Frozen scenario / field ID | Pair, identity, host, surface, and evidence fields | Result | State |
|---|---|---|---|
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r1real-time social sentiment query (<250ms) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=real-time social sentiment query (<250ms); task; response SLA; Grok-4.20 latency; Grok 4.3 latency; SLA compliance; responsiveness winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — real-time social sentiment query (<250ms) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r2breaking financial event analysis (<500ms) | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=breaking financial event analysis (<500ms); task; response SLA; Grok-4.20 latency; Grok 4.3 latency; SLA compliance; responsiveness winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — breaking financial event analysis (<500ms) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r3deep geopolitical crisis synthesis | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=deep geopolitical crisis synthesis; task; response SLA; Grok-4.20 latency; Grok 4.3 latency; SLA compliance; responsiveness winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — deep geopolitical crisis synthesis is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r4continuous background monitoring stream | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=continuous background monitoring stream; task; response SLA; Grok-4.20 latency; Grok 4.3 latency; SLA compliance; responsiveness winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — continuous background monitoring stream is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r5concurrency contention latency spike | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=concurrency contention latency spike; task; response SLA; Grok-4.20 latency; Grok 4.3 latency; SLA compliance; responsiveness winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — concurrency contention latency spike is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch69-grok-4-20-0309-reasoning-vs-grok-4-3-m3-r6unmeasured speed fixture | pair=grok-4.20-0309-reasoning vs grok-4.3; provider=xAI; hostA=api.x.ai; hostB=api.x.ai; surfaceA=xAI API; surfaceB=xAI API; requested/effective IDs=grok-4.20-0309-reasoning,grok-4.3; scenario=unmeasured speed fixture; task; response SLA; Grok-4.20 latency; Grok 4.3 latency; SLA compliance; responsiveness winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — unmeasured speed fixture is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
First-party provenance: All AI Ask measured speed dataset; verification date 2026-09-08. 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 the grok-4-20-0309-reasoning-vs-grok-4-3 Batch 69 scenario →
How do Grok-4.20 Reasoning and Grok 4.3 compare on specs?
| Grok-4.20 Reasoning | Grok 4.3 | |
|---|---|---|
| Price (input) | $2.00/M | $1.25/M ✓ |
| Price (output) | $6.00/M | $2.50/M ✓ |
| Blended price | $3.00/M | $1.56/M ✓ |
| Context window | 1,000,000 tokens | 1,000,000 tokens |
| Max output | 64,000 tokens | 64,000 tokens |
| Modalities | text, vision | text, vision |
| Reasoning mode | Yes | Yes |
| Released | 2026-03 | 2026-06 |
| Speed | 52 t/s | 98 t/s ✓ |
How much do Grok-4.20 Reasoning and Grok 4.3 cost at scale?
| Tokens / month | Grok-4.20 Reasoning | Grok 4.3 | Delta |
|---|---|---|---|
| 1,000,000 | $3.00 | $1.56 | $1.44 (1.9×) |
| 10,000,000 | $30.00 | $15.63 | $14.38 (1.9×) |
| 100,000,000 | $300.00 | $156.25 | $143.75 (1.9×) |
Choose Grok-4.20 Reasoning if…
- ✓State-of-the-art document analysis at 1M context
- ✓Dedicated reasoning mode
- ✓Strong vision understanding
- ✓Long-document analysis and problems that benefit from explicit reasoning.
Choose Grok 4.3 if…
- ✓xAI’s general-purpose flagship
- ✓Extremely fast for its size
- ✓1M-token context
- ✓General-purpose work where speed and a huge context window both matter.
Run this exact matchup right now
Send the same prompt to Grok-4.20 Reasoning and Grok 4.3 side by side and see the outputs yourself.
Try Grok-4.20 Reasoning vs Grok 4.3 FreeWhat are common questions about Grok-4.20 Reasoning and Grok 4.3?
Is Grok-4.20 Reasoning cheaper than Grok 4.3?
Grok 4.3 is cheaper, at $1.56 per million blended tokens vs $3.00 for Grok-4.20 Reasoning.
Which has the bigger context window, Grok-4.20 Reasoning or Grok 4.3?
Both models support 1,000,000 tokens of context.
Can Grok-4.20 Reasoning replace Grok 4.3 for coding?
Both are viable for coding. State-of-the-art document analysis at 1M context (Grok-4.20 Reasoning) vs xAI’s general-purpose flagship (Grok 4.3) — pick based on which strength matters more for your workload.
Which is faster, Grok-4.20 Reasoning or Grok 4.3?
Grok 4.3 is faster: 98 t/s vs 52 t/s, measured on our speed benchmarks.
Neither of these? See Grok-4.20 Reasoning alternatives or Grok 4.3 alternatives.
What related comparisons help choose between Grok-4.20 Reasoning and Grok 4.3?
Grok-4.20 Reasoning pricingGrok 4.3 pricingvs Claude Opus 4.8vs Grok-3vs Claude Opus 4.8vs Claude Sonnet 5Premium model tests
Pricing verified 2026-04-06. Specs verified 2026-08-14.