SWE-1.7
Shares rank 10 of 15 coding models by recommenders
Recommended as a fast, low-cost coding model that comes close to frontier performance on price. Most common caveat: benchmark claims are self-reported and it only works inside Cognition's own harness.
Recommended as a fast, low-cost coding model that comes close to frontier performance on price. Most common caveat: benchmark claims are self-reported and it only works inside Cognition's own harness.
34 people gave an opinion
- recommend it
- 14
- criticize it
- 20
- approval share
- 41%
Recommended in 4 of 13 months
- November 2025: recommended
- December 2025: not recommended
- January 2026: recommended
- February 2026: not recommended
- March 2026: not recommended
- April 2026: not recommended
- May 2026: not recommended
- June 2026: not recommended
- July 2026: recommended
- August 2026: not recommended
- September 2026: recommended
- October 2026: not recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 14 people recommend itNeeds 10 or more, met.
- 0.7 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 4 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Good quality for the price
“I bought the 20$ subscription and have been using SWE-2 for a few days now, and it's actually good.”
Comment from 16 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” “a model from a smaller lab can beat weeks-old Fable 5.1 at 70% lower cost”
Comment from 10 Sep 2026 in “Cognition's SWE-2 achieves 92.8 on Terminal-Bench 2.1” “great that we have a new leader in performance/price in that segment”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” Impressive technical achievement
“This is an excellent write up of an excellent technical achievement.”
Comment from 8 Jul 2026 in “SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence” “It's seriously quite impressive.”
Comment from 11 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” Improving with each release
“windsurf swe-1 is good enough for my use case and swe-1.5 is even better”
Comment from 6 Jan 2026 in “Why didn't AI "join the workforce" in 2025?” “the models did get better but yeah their early product was godawful”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” “SWE-1.5 was surprisingly good when I used it last.”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra”
What people warn about
Each mark is one comment
Benchmarks self-reported or unverified
“A company whose first demo was completely fraudulent announces that its model beats GPT-5.5, on its own benchmark?”
Comment from 8 Jul 2026 in “SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence” “on their own benchmark that they made, that they ran themselves in their own closed-source ecosystem”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” “Seems like benchmaxing? For example for Terminal-Bench 4 it doesn't have great results.”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” - Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra”
Only usable in its own harness
“Would have been worth a consideration if it could have been used beyond it's own harness.”
Comment from 8 Jul 2026 in “SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence” “I'm disappointed to see I need to use a bespoke platform to interact with this agent”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” “they all have desktop apps and have a plan, how do i use SWE-2?”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” - Comment from 8 Jul 2026 in “SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence”
Trails rival models in practice
“1.7 was unusable for me. Took more time thinking than GLM 5.2 and seemed to be generally running in circles.”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” “I'm not using yet another tool and subscription for a model that isn't even winning.”
Comment from 8 Jul 2026 in “SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence” “The model also seems worse in every way than DeepSeek v4.1 Flash, launched today.”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra” - Comment from 10 Sep 2026 in “Cognition's SWE-2 achieves 92.8 on Terminal-Bench 2.1”
Uses up paid quota quickly
“took 14% of my daily quota in one prompt on the $20 plan”
Comment from 8 Jul 2026 in “SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence” “it doesn't appear to be free, it used up some of my on demand usage”
Comment from 10 Sep 2026 in “Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra”
Other coding models people recommend
The four most recommended other coding models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Cursor Composer | 214 | 79 | 73% | 12 | 1 mo11 owners | Approved |
| Qwen3-Coder | 165 | 49 | 77% | 12 | 2 mo7 owners | Approved |
| GPT-5.3-Codex | 135 | 78 | 63% | 11 | 2 mo4 owners | — |
| Meta Muse Glimmer 30B | 70 | 18 | 80% | 3 | — | — |
| SWE-1.7This page | 14 | 20 | 41% | 4 | — | — |
What people said
40 of the 41 comments that mention it, newest first. Each links to the original comment.
“Oddly,* SWE-2 is there for me.”
“I bought the 20$ subscription and have been using SWE-2 for a few days now, and it's actually good.”
“No it still requires Pro in the CLI.”
“It's seriously quite impressive.”
“Cognition has a good frontier model and they are building serious products”
“I don’t have anyone in my life (Devin shop) saying good things about SWE other than it’s free”
“my personal experience with swe has been…suboptimal”
“It’s very far from Fable on benchmarks that matter, like TB4.”
“great that we have a new leader in performance/price in that segment”
“a model from a smaller lab can beat weeks-old Fable 5.1 at 70% lower cost”
“on their own benchmark that they made, that they ran themselves in their own closed-source ecosystem”
“Their Terminal Bench 4 score is soso. Still quite impressive though.”
“i don't fundamentally see the model being better”
“any claimed improvements in performance should be taken with a grain of salt”
“The horrible website is made by Claude or Cognition is distilled.”
“they all have desktop apps and have a plan, how do i use SWE-2?”
“it doesn't appear to be free, it used up some of my on demand usage”
“this is the definition of bench maxxing”
“the models did get better but yeah their early product was godawful”
“Closed weights AND benchmaxxed.”
“1.7 was unusable for me. Took more time thinking than GLM 5.2 and seemed to be generally running in circles.”
“The model also seems worse in every way than DeepSeek v4.1 Flash, launched today.”
“I'm disappointed to see I need to use a bespoke platform to interact with this agent”
“SWE-1.5 was surprisingly good when I used it last.”
“Seems like benchmaxing? For example for Terminal-Bench 4 it doesn't have great results.”
“code that passes tests, but only because they also wrote the tests”
“1.7 tried for 17s before I killed it”
“took 14% of my daily quota in one prompt on the $20 plan”
“when I use swe it feels like massive shit”
“the benchmarks they use are cherry-picked”
“This is an excellent write up of an excellent technical achievement.”
“If I wanted an arranged marriage I could go back to Claude Code.”
“I'm not using yet another tool and subscription for a model that isn't even winning.”
“Would have been worth a consideration if it could have been used beyond it's own harness.”
“Not open source. Also, not available beyond it's own harness.”
“It's pretty good but not Opus 4.8 or Fable levels obviously.”
“Kinda funny that their "cost-vs-performance" chart looks the same as the one for Composer 2.5”
“A company whose first demo was completely fraudulent announces that its model beats GPT-5.5, on its own benchmark?”
“windsurf swe-1 is good enough for my use case and swe-1.5 is even better”
“Takeaway is that this is sonnet-ish model at 10x the speed.”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.