GPT-5.6 Terra vs GLM-5.3

head-to-head

what do the benchmarks actually say?

On SWE-bench Verified — real, human-validated GitHub issues resolved end-to-end — GPT-5.6 Terra posts 95.4% against 95.4% for GLM-5.3, a 0-point gap. Verified is the closest public proxy for "can it fix a real bug in a real repo without help", which is why it anchors our ranking.

A few points either way is real but not decisive: within that band, the agent scaffolding around the model — how it retrieves files, runs tests, and retries — often matters as much as the base model. Treat the gap as a lean, not a verdict.

which is cheaper to run?

Public per-token pricing isn't confirmed for both models, so we don't print a cost comparison yet.

when to pick each

how were these scores verified?

We only print a number once it's confirmed against a primary source or an independent evaluation, and each row on our leaderboard records which kind it is:

• GPT-5.6 Terra: Independent (vals.ai, re-evaluated, mini-swe-agent bash-only harness): SWE-bench Verified 95.4% ±0.94, 5th of 86 systems. CORRECTION, 2026-08-20: this row previously carried 75.2% ±1.93 (32nd of 75), the figure vals.ai published on Jul 22 2026 and which we added Jul 27, 2026. A routine sweep found vals.ai had re-run the model and revised the score up by 20.2 points; the old figure and the reading we built on it, that Terra was the weakest of the three GPT-5.6 tiers, are both withdrawn. A jump that large on an unchanged model points at the first run rather than the model, so read 95.4% as the corrected measurement and not as an improvement. Terra now sits second among the GPT-5.6 tiers, behind Sol at 96.2% and ahead of Luna at 93.0%. Treat it as tied with GLM-5.3 (also 95.4% ±0.94), Grok 4.6 (95.6%) and Claude Fable 5 (95.0%): pooled SEM across those rows is about 1.3 points and every gap is smaller than that. Cheapest of the 95%-plus cluster on measured cost per test at $0.40, against $1.15 for Sol and $2.05 for Fable 5. Repriced Aug 1, 2026: OpenAI cut Terra 20% on Jul 30, 2026, from $2.50/$15 to $2/$12 per 1M (vendor-announced, reported by BleepingComputer and VentureBeat). Cached input $0.25 at launch, not re-confirmed after the cut. Context ~1.05M, max output 128k. Released Jul 9, 2026.
• GLM-5.3: Independent (vals.ai, 2026-08-20 sweep, mini-swe-agent bash-only harness): SWE-bench Verified 95.4% ±0.94, 6th of 86 systems. Left the verifying queue on 2026-08-20, six days after launch. Z.ai published NO SWE-bench Verified figure of its own, so this rank rests entirely on the independent run, which is the cleanest kind of row on this board. Read it as tied with GPT-5.6 Terra (95.4% ±0.94), Grok 4.6 (95.6%) and Claude Fable 5 (95.0%); pooled SEM is about 1.3 points and every gap is inside it. That result is striking given the release is post-training only, on the same base model as GLM 5.2, which is ranked far below on an independent 82.8%. Vendor numbers from the Aug 14 2026 release post, run mostly inside a Claude Code 2.1.207 harness at max effort rather than a neutral one: Terminal-Bench 3.0 28.3 (up from GLM-5.2’s 4.6), DeepSWE v1.1 66.9 (from 46.2), Agents’ Last Exam 28.5 (from 23.8). Its headline claims were cyber rather than coding: CyberGym 84.5% and ExploitBench 54.4%. Marked closed because the weights are NOT out: Z.ai said roughly two weeks after launch pending safety hardening, with no license announced, so the open-weights flag flips only when they actually land. See /p/glm-5-3-cybergym-open-weights-delayed/.

Full reviewsGLM-5.3, decoded

Ranked on our AI Coding Leaderboard, updated 2026-08-20. Scores are confirmed against primary sources; prices are per 1M input tokens and can change.