DeepSeek V4 Flash (Workers AI) vs GPT-6 Luna
Reading a 26.2K-token pull request costs $0.0115 on DeepSeek V4 Flash (Workers AI), 4.4× the $0.0026 it costs on GPT-6 Luna. Both sit in the economy price tier, so the choice between them is about review quality more than budget. The benchmark cannot separate their recall: DeepSeek V4 Flash (Workers AI) minus GPT-6 Luna is −19 pts (95% interval −33 pts to −5 pts).
Side by side
| Figure | DeepSeek V4 Flash (Workers AI) | GPT-6 Luna |
|---|---|---|
| Price tier | Economy | Economy |
| Bugs caughtshare of known bugs | 49% | 67% (better) |
| False alarmsper clean PR | 1.3 | 0.7 (better) |
| Cost per reviewmeasured | $0.0178 | $0.0028 (better) |
| Cost per bug caught | $0.0309 | $0.0046 (better) |
| Exact linecatches on the expected lines | 88% | 97% (better) |
| Review timemean | 137s | 29s (better) |
| Typical review calllist price | $0.0147 | $0.0038 (better) |
| Input / 1M tokens | $0.44 | $0.10 (better) |
| Output / 1M tokens | $1.32 | $0.50 (better) |
| Context window | 1M | 1.1M |
Head to head
The known bugs both were scored on, by who caught each in at least one run.
By language
| Model | All | TypeScript6 bugs | Go4 bugs | Java4 bugs | Rust4 bugs | C#3 bugs | JavaScript3 bugs | Kotlin3 bugs | PHP3 bugs | Python3 bugs | Ruby3 bugs | Swift3 bugs | C++2 bugs | C1 bug |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V4 Flash (Workers AI) | 49% | 46% | 25% | 55% | 46% | 22% | 56% | 61% | 78% | 61% | 44% | 67% | 50% | 0% |
| GPT-6 Luna | 67% | 83% | 25% | 65% | 75% | 56% | 67% | 67% | 67% | 100% | 67% | 100% | 50% | 0% |
Of the 42 known bugs both were scored on, both caught 23 and neither caught 6. DeepSeek V4 Flash (Workers AI) caught 7 bugs that GPT-6 Luna missed: 2 in Go, 1 in Java, 1 in Kotlin, 1 in PHP, 1 in Ruby and 1 in TypeScript. GPT-6 Luna caught 6 bugs that DeepSeek V4 Flash (Workers AI) missed: 3 in TypeScript, 1 in C#, 1 in Go and 1 in Ruby.
On pull requests with no known bug, DeepSeek V4 Flash (Workers AI) raised 1.3 false alarms per review and GPT-6 Luna 0.7. A measured review cost $0.0178 on DeepSeek V4 Flash (Workers AI), 6.3× the cost of $0.0028 on GPT-6 Luna.
| Pull request | Language · kind | DeepSeek V4 Flash (Workers AI) | GPT-6 Luna |
|---|---|---|---|
| golang/gocrypto/tls: add LocalCertificate to ConnectionState | Go · Logic | 3/4 | 0/1 |
| withastro/astroPort Go's csrf protection | TypeScript · Logic | 3/4 | 0/1 |
| socketry/asyncBarrier wait in completion order + `Queue#close`. | Ruby · Concurrency | 2/3 | 0/1 |
| jhy/jsoupStore cached child els in Attributes, not field | Java · Concurrency | 2/6 | 0/2 |
| kotest/kotestFix DataClassEq comparator | Kotlin · Logic | 1/3 | 0/1 |
| symfony/symfony[Security] Allow multiple OIDC discovery endpoints | PHP · Security | 1/3 | 0/1 |
| traefik/traefikAdd the option to define custom cipher suites for backend serversTransport | Go · API misuse | 1/4 | 0/1 |
| OrchardCMS/OrchardCoreSimplify using `StringValues` in Liquid templates | C# · Security | 0/3 | 1/1 |
| TanStack/queryfix(query-core): fix hydration bugs for already resolved promises | TypeScript · Logic | 0/4 | 1/1 |
| colinhacks/zodfeat(v4): parse input containing reference cycles | TypeScript · Resource handling | 0/4 | 1/1 |
| go-chi/chifeat: middleware.ClientIP, a replacement for middleware.RealIP | Go · Security | 0/3 | 1/1 |
| rails/railsReimplement Action Cable redis adapter with redis-client | Ruby · Logic | 0/3 | 1/1 |
| trpc/trpcfix(server): pass maxDurationMs timeout signal to subscription handlers | TypeScript · Concurrency | 0/1 | 1/1 |
Defects caught over all runs: a PR reviewed three times with one known bug shows 3/3 when it was caught every time.
Every bug each was scored on: DeepSeek V4 Flash (Workers AI) and GPT-6 Luna.