Claude Haiku 5.5 vs GLM-5.3 (Workers AI)
Reading a 26.2K-token pull request costs $0.0367 on GLM-5.3 (Workers AI), 14× the $0.0026 it costs on Claude Haiku 5.5. Claude Haiku 5.5 is a economy model and GLM-5.3 (Workers AI) a mid-range one. The benchmark cannot separate their recall: Claude Haiku 5.5 minus GLM-5.3 (Workers AI) is +3 pts (95% interval −11 pts to +16 pts).
Side by side
| Figure | Claude Haiku 5.5 | GLM-5.3 (Workers AI) |
|---|---|---|
| Price tier | Economy | Mid-range |
| Bugs caughtshare of known bugs | 64% (better) | 61% |
| False alarmsper clean PR | 1.4 (better) | 1.9 |
| Cost per reviewmeasured | $0.0038 (better) | $0.0311 |
| Cost per bug caught | $0.0064 (better) | $0.0559 |
| Exact linecatches on the expected lines | 82% | 85% (better) |
| Review timemean | 15s (better) | 65s |
| Typical review calllist price | $0.0038 (better) | $0.0473 |
| Input / 1M tokens | $0.10 (better) | $1.40 |
| Output / 1M tokens | $0.50 (better) | $4.40 |
| Context window | 1M | 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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Claude Haiku 5.5 | 64% | 83% | 75% | 40% | 75% | 67% | 33% | 67% | 100% | 39% | 67% | 67% | 50% | 0% |
| GLM-5.3 (Workers AI) | 61% | 78% | 19% | 46% | 73% | 48% | 33% | 72% | 44% | 58% | 67% | 100% | 100% | 67% |
Of the 42 known bugs both were scored on, both caught 26 and neither caught 4. Claude Haiku 5.5 caught 2 bugs that GLM-5.3 (Workers AI) missed: 1 in Go and 1 in PHP. GLM-5.3 (Workers AI) caught 10 bugs that Claude Haiku 5.5 missed: 2 in Java, 1 in C, 1 in C++, 1 in JavaScript, 1 in Kotlin, 1 in Python, 1 in Rust, 1 in Swift and 1 in TypeScript.
On pull requests with no known bug, Claude Haiku 5.5 raised 1.4 false alarms per review and GLM-5.3 (Workers AI) 1.9. A measured review cost $0.0038 on Claude Haiku 5.5, 88% less than $0.0311 on GLM-5.3 (Workers AI).
| Pull request | Language · kind | Claude Haiku 5.5 | GLM-5.3 (Workers AI) |
|---|---|---|---|
| go-gitea/giteaperf(gitdiff): optimize inline diff highlighting using cache | Go · Error handling | 1/1 | 0/3 |
| symfony/symfony[Security] Allow multiple OIDC discovery endpoints | PHP · Security | 1/1 | 0/3 |
| TanStack/queryfix(query-core): fix hydration bugs for already resolved promises | TypeScript · Logic | 0/1 | 3/4 |
| aio-libs/aiohttpkeep websocket compression state across interleaved control frames | Python · Logic | 0/1 | 1/4 |
| c-ares/c-aresEvent Thread could stall when not notified of new queries on existing connections that are in a bad state | C · Resource handling | 0/1 | 2/3 |
| ggml-org/llama.cpprpc : cache and reuse compute graphs | C++ · Security | 0/1 | 3/3 |
| jhy/jsoupStore cached child els in Attributes, not field | Java · Concurrency | 0/2 | 2/6 |
| kotest/kotestFix DataClassEq comparator | Kotlin · Logic | 0/1 | 2/3 |
| mockito/mockitoAdd core API to enable Kotlin singleton mocking | Java · API misuse | 0/1 | 1/4 |
| nushell/nushellreorganize nushell configuration | Rust · Error handling | 0/1 | 1/4 |
| pinojs/thread-streamrefactor: replace setTimeout with Atomics.waitAsync | JavaScript · Concurrency | 0/1 | 1/3 |
| pointfreeco/swift-navigationUIKit: Fix switch presented child double dismiss | Swift · Logic | 0/1 | 3/3 |
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: Claude Haiku 5.5 and GLM-5.3 (Workers AI).