ModelBrief · Gemini 4 Carbon · Topic Page
Last updated Nov 13, 2026

Quantslant's 50+ Task Test: Carbon vs Argon, Fable 5.2 and GPT-6.1 Sol — Coding Is the Weak Spot

Published Nov 13, 2026 · Last updated Nov 13, 2026

Quantslant tested Carbon twice right after GA: a first test within an hour of availability, then a 50+ task shootout against Argon, Claude Fable 5.2 and GPT-6.1 Sol. Everything ran through OpenRouter with identical settings and was graded by unit tests and answer keys — the most methodical hands-on testing available so far. The headline: Carbon aced nearly everything on everyday and agent work, but coding is a clear weak spot — 63 of 76 unit tests, where all three rivals scored 76/76.

Everyday tasks: tied with Argon, twice as slow

23 practical tasks (an agent with tools, code with unit tests, writing and proofreading, puzzles, a 158,000-character document), at low effort:

ModelScoreCostTime
Gemini 4 Carbon99.5%$0.23264s
Gemini 4 Argon99.5%$0.22135s
Claude Fable 5.2100%$1.37190s
GPT-6.1 Sol100%$0.22180s

Carbon's missing 0.5% was a grader error — in practice it matched Argon exactly, in twice the time. The tester's verdict was blunt: on everyday tasks, Carbon does nothing Argon cannot do.

Agent, spreadsheet and security: all perfect

Agent task (6 documents plus a trap): all four models 10/10, Carbon slowest at 54s. Spreadsheet engine, 34 unit tests: all 34/34, but Carbon took 518 seconds and $0.47, writing 639 lines including its own parser and generator-based evaluation, against Argon's 64 seconds and $0.05. Security review of a small Flask app with 8 planted vulnerabilities: Carbon found all of them with correct lines and CWE numbers, in 67 seconds for $0.03. SOC log analysis: Carbon 7/7, identifying the brute-force IP, the compromised account, the exact second of the login, a CSV export 20 seconds later, and a URL-encoded ..%2f path traversal.

The weak spot: coding, 63/76

On 6 coding tasks with 76 unit tests (medium effort) — the only part of the shootout where Carbon clearly lagged — Carbon passed 63/76 while Argon, Fable 5.2 and Sol all passed 76/76. That sits oddly with Google's official table, where Carbon leads Terminal-Bench 4.0 at 65.2%: official benchmarks and these unit-tested practical coding tasks are not measuring the same skill. The sample is still small (6 tasks), but the direction of the gap is worth noting: anyone buying Carbon as "the strongest coding model" should first test it on their own real repository.

Made-up sources and vision: Carbon is honest

Asked for 8 scientific papers with DOIs on two topics, Carbon listed 16 publications that all exist, with only two minor errors (one wrong DOI, one wrong year) and zero inventions; Argon invented 2; Fable 5.2 had 0 inventions and 0 errors; Sol 0 inventions and 1 minor error. Vision tasks: 10/10. This matches the AA-Omniscience measurement: Carbon's hallucination rate (11%) is higher than Argon's (4%), but its tendency to fabricate an answer instead of admitting ignorance is low at medium effort.

The thinking-burn bug: tokens spent, answer empty

It happened twice. In the first test, with a 6,000-token output cap, Carbon spent 5,808 tokens on reasoning and returned an empty answer — still billed. In the shootout, a regex task at medium effort used all 16,000 tokens on reasoning and returned no code ($0.16). This is a mechanical consequence of the output cap covering thinking tokens, not a random glitch — Google's docs give the same advice: control cost with the thinking level, not a low output cap. See our thinking levels and pitfalls guide.

Our take: Quantslant's testing paints a clear picture of Carbon: no better than Argon on everyday work, and slower; strong on "search, check, verify" work like agents, security and log analysis; but behind all three rivals on practical coding. Carbon is not a general upgrade — it is a specialist tool, worth its time and token bill only on long, hard tasks that need repeated verification.

Reference source: this article is an original rewrite produced by this site; facts and figures were checked against Quantslant 50+ task four-model test and first test (Nov 11, 2026)

← Back to the Gemini 4 Carbon topic page