Code SmithOperatorPeer public scoresmeth_v0_1_1as of 2026-07-11

Grok 4.5 · API · High effort

The alt-frontier code smith — near-Opus agentic coding at a third the token bill.

SpaceXAI Grok 4.5 for fast, cheap frontier coding and terminal work when you want Claude/GPT rivals without $25–30 output rates.

SharePost on X

Codex rating

74.0

Confidence B

Overall mean of attribute public scores after peer min-max (40–95 band within scored launch peer set).

Loadout

Endpoint: grok-4.5
Product: SpaceXAI / xAI API
Price: $2 / $6 per MTok
Context: 500k tokens

Grok 4.5 API at high reasoning, 500k context, $2/$6 per MTok. Alt-frontier coding and agentic Build — strong cost/tempo story vs Claude Opus and GPT Sol.

Core attributes

Six stats · peer public

Peer publicbars = rank in scored peer set (≈40–95) · raw = absolute fusion under the bar

STRExecution

68.8raw 81.7A64.8–72.8

DEXTempo

76.7raw 68.0B66.7–86.7

CONReliability

65.6raw 81.3A61.6–69.6

INTReasoning

71.8raw 84.2A67.8–75.8

WISJudgment

72.1raw 70.7A68.1–76.1

CHACollaboration

88.9raw 89.0B78.9–98.9

Peer public (main number / bar) ranks this Build inside the current scored peer set (≈40–95). Raw is the absolute evidence-fusion aggregate before min-max. A low peer score can still be a strong absolute model — see methodology.

Practical verdict

When to choose

Best at

Cost-efficient agentic coding (named harnesses)
Terminal / multi-step engineering loops
Greenfield apps and high-tempo iteration

Avoid for

EU-only deployments until regional GA
Trust-critical judgment without verification (54% Omniscience hall)
Maximum absolute coding ceiling at any cost (Sol / Fable still lead Verified)

Choose this when

You want near-frontier coding and agentic work with strong price and speed.

Skills

Task proficiencies

Creative Writing89.0 B

Professional Writing89.0 B

Test Writing86.6 A

Devops And Terminal83.3 B

Tool Calling83.3 B

Refactoring82.5 A

Debugging81.7 A

Existing Codebase Work80.5 A

Quantitative Reasoning78.7 A

Following Complex Instructions78.1 A

Long Autonomous Runs76.4 B

Research Synthesis75.2 A

Fast Interactive Work68.0 B

Traits & Quirks

Signature strengths and honest flaws

No structured traits or flaws catalogued for this Build yet. Traits are behavioural patterns (beyond raw scores) — added as evidence and field observation accumulate. See Best at / Avoid for above in the meantime.

Evidence Terminal

Traceability

Every factual claim cites an evidence row. Tap a chip to see the benchmark, exact config, source, and caveats behind the number.

GPQA Diamond93.1%A Humanity's Last Exam40.3%A AA-Omniscience26A Artificial Analysis API latency / throughput93.2B Chatbot Arena / LMArena human preference1505B SWE-bench Pro64.7%B SWE-bench Verified86.6%A Terminal-Bench83.3%B

Engine pass meth_v0_1_1 (peer min-max vs scored launch peer set). Codex overall 74 (B). Attribute public scores map raw aggregates into a 40–95 band within the peer set (methodology min-max). INT uses hierarchical fusion (GPQA primary / HLE secondary / capped bleed).