GLM-5.3 Matches Claude Fable 5 on DeepSWE at 25% of the Cost
zainhas · x · 2026-08-23
Together Compute analyzed GLM-5.3 and Claude Fable 5 on the DeepSWE benchmark. GLM-5.3 matched Fable 5 in single-shot solve rate and pulled ahead with multiple attempts.
Key breakdown:
- Solve Rate: GLM-5.3 (69.0%) vs. Fable 5 (69.7%)
- Cost: GLM-5.3 ($3.99/task) vs. Fable 5 ($21/task)
- Tokens: GLM-5.3 (80k) vs. Fable 5 (114k)
- Turns: GLM-5.3 (124) vs. Fable 5 (85)
The analysis highlights that the economics become highly favorable when retries are cheap.
Related event: GLM-5.3 Matches Fable 5 on DeepSWE at a Quarter of the Cost(4 posts)→
More from Models
- Decision models match LLMs on routing and safety — and are 7-8x faster — randal_olson · 2026-10-16
- Xiaomi MiMo V2.6 Pro takes #1 open-model spot on Agent Arena at $0.11/task — arena · 2026-10-16
- ChatGPT's 'best AI video tool' answer points to Runway integration, sparking ad doubts — tlakomy · 2026-10-16
- Microsoft's MAI finishes in-house 3T-parameter pretrained base models, post-training results coming — soumitrashukla9 · 2026-10-16
- Hands-on with Gemini 4 Argon: faster than Opus 5.5 but hallucinates more and costs more — prasenx · 2026-10-16
- Bindu Reddy urges Meta to open-source Wonder, claiming it beats DeepSeek — bindureddy · 2026-10-16